Section 6 - Future Directions

Chapter 25 Modelling in DOHaD Challenges and Opportunities in the Era of Big Data

Julie Nihouarn Sigurdardottir and Salma Ayis

25.1 Introduction

Public health data available for research are booming with the expansion of Big Data sources, shifting the landscape of DOHaD research. These new forms of data offer ample opportunities to advance epidemiological modelling within the DOHaD framework. Big Data is often described by the ‘3 Vs’: high volume, high velocity, and wide variety and refers, for example, to the large volumes of Electronic Health Records (EHRs) now stored as many nations move towards the routine electronic recording and centralising of health data. The term Big Data also applies to data derived from wearable devices and phone applications, increasingly affordable technologies that allow for the collection of new kinds of data, in larger volumes, and almost in real time. Such technologies, along with improved data processing speed and advanced computing capacity, grant access to the lifestyle and health information of millions of individuals who can be followed through the lifespan.

However, within heterogeneous and dynamic socio-demographic contexts and a fast-moving technological landscape, these new forms of data raise a plethora of methodological challenges related to accurately characterising population health trajectories and biological mechanisms. In addition, while the current inferential potential of DOHaD research depends on which variables are collected, at what frequency, and at what time points, it is also closely shaped by the theoretical model(s) chosen for a given study: a framework implicating critical and sensitive windows in development shaped the early DOHaD literature, but other models were added such as the accumulation of risk model, the chain of risk model, and a hybrid of those [1]. These frameworks shape study designs, data collection practices, and the interpretation of results and set the scene for how Big Data is likely to be taken up in the field.

In this chapter, we provide an overview of DOHaD modelling methods and consider the emerging place of Big Data in investigating multidimensional research questions in the field. To do so, we discuss various methodological aspects of modelling, such as operationalisation, sampling, population representation, ethics, and the accuracy of tools used to acquire and analyse data. We also discuss the current landscape of artificial intelligence-derived methods, judging their utility against the validity of findings, and their potential when compared to ‘traditional’ empirical data sources and analytical approaches.

25.2 Current DOHaD Modelling and Methodological Challenges

A myriad of methodological challenges are present in DOHaD research even prior to Big Data, in particular, the issues of validity across time and space, characterising causal links, and identifying sources of bias. Physiological processes are difficult to model because they are integrated into non-static systems. These refer to the less quantifiable and less predictable behavioural, lifestyle, environmental, and socio-economic systems interacting with biology. The moderating or mediating effects of culture, health inequalities, medical systems, and health policy on health outcomes must be clarified to assess the generalisability of any given model. Even in gold-standard birth cohort research designs, epidemiological models need to account for the fact that, for example, through societal restructuring and climate change, the properties of exposures can change over time and across generations [2]. As a result, it is difficult to produce predictive DOHaD models and interventions that remain valid and useful across time for a given population. Alternative designs such as the observational study design in humans cannot, however, capture the complexity of all important causal links.

A further challenge for aetiological and epidemiological models of DOHaD is how to define the sources of individual differences in health outcomes with a robust degree of certainty. Some examples include disentangling antenatal and postnatal exposure and their interactions [3]; accounting for sex and gender-based differences in biology and behaviour; inter-organ variation in adaptability to maternal ill health (e.g. the placenta response to stress) [4, 5]; and evaluating disparities in outcomes across diverse groups given that the bulk of data is from a small number of middle- to high-income and white-dominated contexts [6, 7].

Prediction models are also prone to confounding and collider bias [8] (a variable in a causal pathway that is a shared effect of more than one cause). For DOHaD research, the primary exposures studied in the developmental pathway are nutrition, parental physiological and psychological health, the environment and toxicants, and social and demographic determinants [7]. Intuitively it is easy to assume that many of these exposures can co-occur and may moderate one another. While confounding can often be resolved by taking these variables into account, residual confounding remains a risk when those influencing factors are unknown or unmeasurable. In the case of collider bias, this can also lead to counter-intuitive conclusions. Such counter-intuitive conclusions are exemplified by the ‘birthweight paradox’, where babies born with low birthweight (LBW) to smoking mothers (exposure) appear to have a lower risk of neonatal mortality (outcome) compared to those born with LBW to non-smoking mothers [9].

Here, tools such as directed acyclic graphs (DAGs) that portray causal relationships graphically can help a researcher explore a model’s functional assumptions and conceptualise any mechanisms of causality. DAGs help recognise mediators, moderators, confounders, and colliders [10] and have helped to illustrate the collider role of LBW in the above paradox [9]; that is, when maternal smoking is absent, other unobserved causes (malnutrition and congenital defects) can lead to LBW and more severe health problems and thus higher infant mortality. Therefore, while the above research question appeared ‘simple’ initially and involves few measurable exposures (smoking or not) and outcomes (LBW and mortality), failing to incorporate inter-correlations between variables and confounding effects, conceptualised by theory and DAGs, is unlikely to provide reliable causal inference.

Taking another example, understanding the association between maternal stress and lower infant cognitive outcome [11] warrants pertinent exploration into the relative contributions of other exposures concomitant to maternal stress, such as under- or overnutrition, infections and toxicants, and their interrelationships. Here, modelling methods should integrate observed variables, latent variables (not directly observed but derived from questionnaires or other observed variables, i.e. stress), and their measurement errors, alongside time indicators. Moreover, the mediating mechanisms proposed in the literature, such as epigenetic modulation, the microbiome, metabolism, and (offspring) endogenous immunity, would also have to be incorporated. In practice, a single model that simultaneously incorporates multiple predictive pathways and associations will more accurately capture the causal relationships of interest between the exposure and outcome of interest [12, 13]. Of translational value, such epidemiological modelling would eventually result in developing better targeted interventions.

25.2.1 Data: What Are We Collecting, What Are We Measuring?

High-quality data that are fit for purpose and meet the criteria of accuracy, validity, completeness, and consistency are a cornerstone of empirical science. Data quality may be affected at the stages of data collection, cleaning, or the numerical transformation that is often used to meet required assumptions such as the normal distribution in statistics. Measurement errors, whether systematic or random, are present in all observational studies. While these errors impact the validity and reliability of data and introduce biases, they are rarely acknowledged or accounted for in the epidemiological literature. Makin and de Xivry address common statistical mistakes [14], and Wagenmakers et al. [15] present guidelines on how to report statistical analyses transparently based on four scientific norms of ‘communalism, universalism, disinterestedness and organised skepticism’.

Since we allude above to the notion of variable choice and availability, next, we discuss the importance of clear terminology and data quality in DOHaD research methods. How we define and measure exposures and outcomes impacts inferences, findings, and subsequent interventions and policies. One example of this is the work of researchers who rely on clinically defined groupings based on dichotomisation, such as diabetes diagnosis, or the classification of body mass index (BMI) as obese/overweight/normal-weight/underweight. One obvious risk of using a strict classification of body morphology by BMI alone is undermining the field’s knowledge about fat distribution being a strong determinant for metabolism and cardiovascular health, especially fat within the abdomen (visceral adiposity). Without other markers to corroborate metabolic health risks (blood pressure, cholesterol, visceral fat mass, etc.), some individuals with normal weight, categorised as ‘controls’, may be metabolically unhealthy and ‘at-risk’ of physiological phenotypes. In fact, this group represents 35 per cent of normal-weight individuals [16]. This problem extends to gestational diabetes screening in pregnancy, which is provided to women meeting the BMI > 30 kg/m² criteria in the UK, while those under 30 are assumed to be void of any hyperglycaemia risks during pregnancy. The absence of evidence, however, is not evidence of absence. The consequence of such hidden (latent) subgroups of individuals within a ‘control’ or ‘normal-weight’ category is the introduction of bias into the statistical analyses that epidemiological models rely on for inference, thus leading to inaccurate conclusions.

Similarly, in psychology, the diagnosis of autism as present/absent is common, although autism spectrum disorder (ASD) is typically conceptualised by experts as a continuum (see also Azevedo et al. in this volume). Such a binary diagnosis of ASD ignores potential distinct mechanisms of importance for the DOHaD of autism subtypes, which could possibly relate to the timing of any ‘disruption’ in brain development and could be informative for mechanistic studies and prognosis [17].

Overall, what we suggest above is that DOHaD researchers must be conscious of relying on clinical data alone such as those retrieved from EHRs without clarifying sources of biases, the caveats of present/absent dichotomised diagnoses [18], and the local clinical guidelines from which they derive. We suggest, however, that some classification approaches are available to limit some of these caveats such as collating multiple variables and produce profiles based on similarities of exposure and/or outcomes at one time point (e.g. latent class modelling) or many time points over time to uncover trajectories (latent class growth analysis and piecewise modelling). This could mean, for example, retrieving glucose measures sampled throughout pregnancy and establishing the likely glycaemic status rather than relying only on a single GDM diagnosis. These approaches also help identify profiles of individual responses to interventions and can therefore improve tailored treatment allocation.

Additionally, missing data, either by design (e.g. unmeasured exposures/outcomes) or attrition, negatively impact data quality and challenge causal inference. This can be addressed by powerful analytical tools that recognise data complexity and the impact of missing data [19, 20]. Several methods are available to address missing data, including maximum likelihood, multiple imputation, and Bayesian methods. Complete case analysis leads to loss of data and statistical power but is widely used, while other complex but more justifiable methods are not often attempted [21]. Assumptions about the properties of missing data, whether these are missing completely at random or not, must be made. With the emergence of Big Data and EHRs that tend to have a high prevalence of missing information, appropriate techniques that deal with missing data need to be carefully applied.

25.3 Big Data: Challenges and Opportunities

As referenced above, Big Data refers to data large in volume, collected at high velocity, and comes in a variety of sources, formats, and dimensions, such as from birth cohort and longitudinal studies, medical records, or wearable/phone devices. Birth cohort studies, such as the Avon Longitudinal Study of Parents and Children study, have supported the DOHaD hypothesis by in-depth prospective sampling and large multidimensional data collection from human participants. While integral to the DOHaD evidence base, standard cohort studies are costly and may be of limited size. EHRs, a source of Big Data, can be obtained from centralised systems, while large omics data sets (genomics, transcriptomics, metabolomics, etc.) are often sourced from biobanks and can be added to increasingly available personal and external ‘exposome’ data (e.g. lifestyle and environmental). Among the benefits of EHRs is their level of comprehensiveness, and so with larger samples, this also improves the statistical power required to provide accurate estimates of effect size. The availability of such data means that if taken up in DOHaD research, the scope for such studies would no longer be limited by small sample sizes due to funding and/or the restrictive protocols of conventional longitudinal birth cohorts [22]. In practice already, linkage study designs join primary- and secondary-care databases or merge multiple EHR databases and registries, potentially offering new insight into disease pathways. For example, the UK-based CALIBER study drew on EHR sources to investigate the cumulative incidence and period prevalence of diseases over the lifecourse. Results were presented in the form of a chronological map of 308 physical and mental health conditions from four million individuals, from infants to the elderly [23]. The role of universal medical coverage and centralised digital health records, such as the English National Health Service, in enabling such exploration in this particular study cannot be underestimated.

More recently, data acquired from real-time biosensors measuring pollution exposure, blood glucose levels, or heart rate from wearable devices have become available. Data involving behaviour and social networks can also be retrieved from open social media platforms at high speed. The past three years, especially during the COVID-19 pandemic, have seen a surge of software developments intended to meet the needs of monitoring health markers and well-being remotely. The uptake of telemedicine was enabled, for example, by digital platforms used by clinicians to manage antenatal hyperglycaemia [24] and the self-report of glucose levels by pregnant women on their phones [25]. These tools could be employed in future DOHaD studies. However, key ethical issues related to privacy, rights, and moral code of conduct when retrieving these data require careful considerations in this changing research landscape.

25.3.1 Limitations of Current Big Data Sources and Applications

One first potential caveat of relying on Big Data sources it that DOHaD researchers wanting to use Big Data may be obliged to formulate research questions based on data availability or including data not necessarily designed primarily for DOHaD research. These researchers will have less control over data quality because of the larger distance from data collection, that is the inputting user (clinician for EHR / hospitals, or user of a phone app), and from the decisions made in defining and measuring the variables in these data sets. Overall, sources of error need to be considered when evidence from Big Data is evaluated. Without researchers’ involvement in data collection, it may be impossible to subsequently correct or even identify these errors.

The task of comparing and validating DOHaD models across populations may be further hindered by the heterogeneity in data architectures across national and international sources. Before the term Big Data surfaced, omics-derived data alone (e.g. genomics, transcriptomics, and proteomics) already inferred the outputs of millions of data points [3]. Formulating a cascading model of these omics layers, which follow biologically downstream from one another, is both necessary and extremely complex. Further, linking biologically derived material to clinical data of different formats, and based on a variety of measurements, including imaging, questionnaires, and diagnoses, requires technologies that facilitate multidimensional integration. Thereafter, powerful methods that support analysis are necessary.

Larger sample sizes improve the power to detect effects, and clearly the whole DOHaD framework requires both large samples and a comprehensive set of exposures and events to be modelled. However, the primary issue is that complex models are more difficult to explain and thus could complicate their practical translation into actionable policies. Users and clinicians equally need to be versed in their use and interpretation.

The use of Big Data also raises issues of data security and representation, particularly data obtained outside conventional academic institutions, in contexts where systems and resources are not fit for this purpose, such as in low- and middle-income countries (LMICs). Users of both healthcare services and digital platforms, such as social media, may represent distinct groups with possibly little overlap in demography, risks, and healthcare needs [22]. It is plausible to assume that LMICs are unlikely to have population-wide health records or access to digital data collection and remote health monitoring from which DOHaD modelling could be done. Here, validation and reproducibility of DOHaD models are less feasible and highlight the lack of representation through their exclusion.

25.4 Artificial Intelligence: Challenges and Opportunities

When Big Data is considered, the word AI is not far behind. Pattern recognition, similarity profiling, and predictions are tasks for which AI methods such as machine learning (ML) have been developed, and these have potential applications to DOHaD research. It should be noted that ML and conventional statistical methods may be seen as a continuum since the algorithms behind ML, including linear and logistic regression, and several dimensionality reduction techniques have existed for decades. (For a contrast between ML and conventional statistics, see [26].) However, the real advantage of AI is that it supports the analysis of large data volumes alongside multidimensionality (i.e. where the number of variables is larger than the subjects).

AI is already being tested and implemented in the clinical domain, including to improve the efficiency of hospital administration. AI is also being used to predict medication side effects and patient outcomes from radiological imaging and thereby promote patient-tailored medicine and interventions. In DOHaD, ML approaches would subserve exploratory designs to identify biological pathways, which appear more frequently in the mosaic of data, in the form of associations, including from DNA sequences and omics data, and those obtained from EHRs [26]. Certain applications require user input from which the ML ‘learns’ to classify new data from sets of rules from previous data (supervised learning) or is completely unsupervised in detecting patterns. This is similar to the latent modelling techniques mentioned earlier that derive from the ‘classical’ statistics and the structural equation modelling framework. Other subfields of AI associated with ML include deep learning, rooted in multi-layered neural networks, which also allow computers to identify relations between concepts/features and characterise these associations from complex to simpler concepts [27]. ML methods to date could also assist in the processing of single modalities or data collection methods, such as magnetic resonance imaging of the brain, heart rate variability in the fetus, and DNA methylation patterns in disease [28], all of which are relevant for DOHaD research.

Despite the potential of ML for DOHaD described, very few ML studies have so far transitioned from single data type and scale to ‘fusing’ several dimensions, or and towards the integration of additional outcome measures retrieved from EHRs, medical imaging, and biospecimens. Data harmonising and deployment of ML is an ongoing endeavour, but some attempts have been made in relation to cardiovascular medicine (reviewed by [29]). The issue to date is that these algorithms exploit two modalities at most (e.g. radiological imaging and free text from clinical reports). (For an in-depth review of the current landscape of AI for multi-modal integration, see [30]).

In our previous section, we discussed the necessity to characterise accurately DOHaD prediction models. It is at present the case that the several competing ML approaches and the rapidly evolving demands of AI have yet to produce a consensus regarding how to develop or validate a prediction model relying on these novel tools. For example, in predicting Type 2 diabetes and cardiovascular disease, Dalakleidi et al. reported the best performance to have been achieved by groups of artificial neural networks. However, the so-called decision trees, random forest algorithm, and support vector machines were said to provide the best accuracy measures by Zheng et al. [31]. Furthermore, AI methods do not necessarily outperform conventional statistical regression applications and are not free of methodological biases [32]. Additionally, scholars warn that ML studies are often computationally demanding on resources (e.g. support vector machines, logistic regression, random forests, gradient-boosted machines, and neural networks) [33].

25.4.1 Issues of Interpretation and Reporting

Concerns are often raised about the scope, complexity, transparency, reproducibility across different scientific teams and different populations, and the interpretability of prediction models. While AI operates from a ‘black box’ within deep neural networks and unsupervised learning, the biological plausibility and meanings of the output are generated by researchers. Given that so far only a few published prediction models have found utility in clinical practice, the utility of AI when compared to conventional methods remains an open question. It is unclear whether AI can address the questions of causality most pertinent in DOHaD, when DOHaD draws from interpretability and theory and is moving towards the integration of social science and ethnography. (See Richardson, in this volume for a discussion of how DOHaD is characterised by ‘cryptic causality’.)

Guidelines regarding AI are developing rapidly. For example, following a quality assessment of the conduct and reporting of multi-variable prediction models, a 22-item checklist (TRIPOD) was developed [34]. The risk of bias tool (PROBAST-AI) for diagnostic and prognostic prediction model studies based on AI has also emerged to ensure that users have key information about the design, conduct, and analysis, alongside a robust standardised tool for bias evaluation that would allow a fair judgement on the utility of these models [31, 35, 36]. Guidelines should be consulted by authors, reviewers, and editors, to ensure reproducibility, reliability, and validity, and hence safe implementation. Again, even prior to applying AI-derived analytics, the uncertainty of measurements in Big Data and its sources of error must be accounted for and possibly identified systematically, and the data quality validated.

25.6 Ethical Questions about Big Data and AI

The rapid expansion of Big Data and AI raises a range of ethical concerns. First, the question of who audits and protects data including EHRs (which can also be used as testing data by commercial parties) is central to ensuring ethical research in the DOHaD field and is currently insufficiently addressed [22, 37]. The lack of standardisation of data protection laws between countries adds to this issue.

Additionally, the most powerful AI pipelines are deployed from within the very few corporations with the computational and financial resources. The commercialisation of both healthcare software or AI tools and their findings within the private sector is another challenge that research institutions must navigate if the potential of such data is to be realised. Such a feat would require more transparency and possibly a move to open sources of data. (For an example of Google’s DeepMind approach to open data, see [38].)

Nevertheless, open and public data collection is also likely to introduce other ethical issues that need to be carefully considered [39]. Big Data collection and usage may move the position of the individual (the unique data provider) from the one fulfilling the ‘social vision’ of the healthcare system and science into the ‘economic vision’ of the commercial enterprise [22]. Participants recruited through academic institutions consciously engage with the scientific community with consent and pledge their time voluntarily. This contrasts with the passive and often unwitting involvement in Big Data collection via the data generated by medical records and phone devices. The use of such data without consent raises concerns regarding data ownership, privacy, and the circulation of profits. Data protection in secondary research by academic institutions using public data is enforced by the institutions themselves through university and institutional ethics boards, but enforcing consent and data protection may be less clear when third-party commercial and private bodies are concerned.

Rarely mentioned in the discussion of AI-derived methods are the risks introduced or heightened to certain populations because of their implementation. Experts such as Professor Kate Crawford at the AI Now Institute are re-evaluating the societal burden of AI. She states AI is not ‘artificial’ since it requires the same earthly resources and labour to mine the power and hardware sustaining it. Consequently, this is also becoming a source of disparity and power imbalance on the ground, within and between populations who compete to mine these resources [40]. Such ramifications would be the real irony for an AI integration into DOHaD research and the long-term agenda of the scientific community.

25.7 Conclusion

There is a strong anticipation that in the future, DOHaD researchers will benefit from innovative methodological designs. This could build on the best of current biostatistical methods and soon include AI technologies where the multidimensionality of data sources and a longitudinal format can be integrated, and the outputs shown to be interpretable. Current and novel ‘mega’ projects may push this progress forward. An example is the protocol implemented in the EarlyCause project [41] that will explore the causal mechanisms between early-life adversity (antenatal and postnatal) and future psycho-cardio-metabolic multi-morbidity. It will involve the participation of 14 European institutions and include three complementary and sequential phases that integrate longitudinal population data sets (e.g. ALSPAC and UK BIOBANK), animal studies, and cellular models with analytical tools from structural equation modelling and machine learning. It also aims to offer a web-based platform for data access and information on research standards and best practices to support future study designs and exploration. Such a mix of granular data collection and Big Data sources in open access, along with AI and conventional statistical approaches, holds great potential for DOHaD research.

The DOHaD research community may look to other fields and consider how to train their own data and solution architects in the newest technologies and Big Data usage. Interdisciplinary teamwork will be crucial in ensuring both robust management and use of data as well as anticipating ethical and governance issues [42]. It is crucial to assess whether certain limitations are inevitable or can be remedied to create the necessary, transparent, and reliable evidence base. Collaborations in data collection could be expanded more frequently to ‘crowdsourcing’ in data analysis and interpretation. Of course, teamwork does not come without caveats when studying complex and dynamic modelling, such as leading to further heterogeneity in findings and conclusions [43]. Nevertheless, we reiterate that attention to operationalisation of exposures and outcomes, reducing bias in data collection and analysis, and the necessity for interpretability should be at the forefront of the DOHaD agenda in the era of Big Data and AI.

Chapter 26 The Promise of Reversibility in Neuroepigenetics Research on Traumatic Memories

Stephanie Lloyd , Pierre-Eric Lutz , and Chani Bonventre

26.1 Introduction

Just over 20 years ago, molecular biologists Leonie Ringrose and Renato Paro published an article with a provocative title: ‘Remembering Silence’ [1]. The article focused on how epigenetic elements, modified through a variety of means, could subsequently return to their silent state. Silencing is operationally defined as their epigenetic status before modulation by experimental or environmental factors. Ringrose and Paro’s article described research on fruit flies and factors affecting embryological growth. Yet it asked a question of considerable importance to parallel and rapidly expanding research in human neuroepigenetics, that of reversibility of the molecular impact of the environment on an individual’s biological profile. In the case of epigenetic modifications that are thought to be mediators between life trauma and the risk of psychopathology, this question would be translated as follows: if you experience a traumatic event and, as a result, acquire an epigenetic state considered to place you at higher risk, can you free yourself of that state? Through a critical assessment of contemporary neuroepigenetics research, in this chapter we consider researchers’ ambitions to account for the indeterminacy of life and the speculative possibility of reversing acquired epigenetic states. Bringing together the perspectives of medical anthropology and molecular biology, we are interested in clarifying how reversibility – a return to silence – is envisioned, how therapeutic interventions purported to bring about that silence might function, and what this might mean for the mental health of people who live in the aftermath of trauma.

The question of reversibility is compelling for a wide range of research agendas in epigenetics, a science that has produced an evidentiary base of significant importance for the field of Developmental Origins of Health and Disease (DOHaD). Indeed, epigenetics research has provided insights into the molecular means by which life experiences might be associated with risk and resilience for the subsequent development of pathology. While the concepts of risk and resilience have received increasing attention in developmental research in recent years, little is known about their purportedly associated epigenetic states, such as their durability. In neuroepigenetics research, resilience is conceived of as a mechanism that may recruit different biological pathways than those triggered by adversity. It is often assessed in two ways: behaviourally and molecularly. Behavioural resilience has been conceived of as the possibility of not being affected by a negative experience at psychological and clinical levels; in other words, as being able to actively counteract what are considered pathological molecular states. Molecular resilience has been studied as the failure to be negatively affected, in terms of acquired epigenetic states, by adverse circumstances. The two conceptualisations of resilience are drawn together in experimental contexts, most often with model organisms, in which molecular profiles are sought to explain why different animals might exhibit what are seen as at-risk or resilient states. Risk, for its part, has been framed as the development of an epigenetic state associated with psychopathology, following a traumatic event; in effect, a molecular memory of that event.

Yet neuroepigenetics researchers only have speculative models to guide studies of the type or number of epigenetic states considered sufficient to confer risk or resilience in the face of adverse experiences, alongside the means by which one might reverse acquired risk. Efforts at reversibility – or remembering silence – by necessity include considerations of the relationship between subjective states, past events, and memories of those events. Since past events cannot change, it is the memory of these experiences that may be the target of a panoply of clinical evaluations and interventions (whether pharmaco- or psychotherapy). Neuroepigeneticists consider mapping these processes an urgent priority given the prevalence of trauma; for instance, approximately 80 per cent of the American population is thought to have experienced trauma-level events [2]. These statistics are deemed particularly worrying given research that suggests that the epigenetic effects of traumatic events may contribute to a variety of pathologies, from cardiovascular to suicide risk, including anxiety and depressive disorders, addiction, and more [3].

Researchers hope that a greater understanding of molecular memories (i.e. epigenetic states) thought to be acquired through the experience of traumatic events, and their relationship to subsequent risk of psychopathology, might allow the development of targeted interventions to help people ‘remember silence’: to reverse the effects of presumably acquired pathological traits. While pre-existing and emerging models alike tend to presume the durability of epigenetic states acquired during early life, Ringrose and Paro’s evaluation of epigenetics research remains productively provocative as it conceptually fuels the hypothesis of the potential for epigenetic reversibility. It also foreshadows more recent shifts in some DOHaD research agendas that are moving away from deterministic models of early-life experiences leading to diseases later in life and are instead focused on conditioning, which implies the possibility for change [4].

Through the following sections, we discuss polysemic understandings of memory and how research on reversibility is entangled with metaphors of silence as a subjectively untroubled or unaffected state. We begin with a consideration of the tensions between narratives of reversibility and persistence in epigenetics research to sketch out what is currently known and unknown about these processes.

26.2 Persistence and Reversibility in Epigenetics Research

In 2001, biologists Ringrose and Paro evaluated emerging research, indicating that, in Drosophila, regulatory elements that are experimentally switched to their active state can ‘“remember” and restore their previous [silent] state’. These ‘regulatory elements’ are defined as regions within genes where, under epigenetic control, proteins that regulate gene activity may have different functional impacts. The authors noted that silenced states can be remembered after several cell generations during which those elements were active, though they could only hypothesise as to how or why regulatory factors would return to silence. This article dates from the early days of epigenetics research, yet Ringrose and Paro’s interests in how epigenetic elements change, with what effects, and whether they are reversible persist. In their contemporary version, they might be: how might epigenetic states acquired through exposures contribute to health and disease? And are the molecular traces of those experiences reversible?

The research discussed by Ringrose and Paro yielded findings on the varying effects of single epigenetic alterations depending on the type and timing of the modification. Each of the studies raised questions about the stability and reversibility of epigenetic states and their developmental effects. For instance, even if an epigenetic state is only modified for a limited period of time, it will nonetheless affect downstream biological processes, which may have longer term consequences than the bout of epigenetic plasticity itself. Ringrose and Paro also observed that while certain experimental data suggested that the restoration of silence was not possible after a significant period of activation, other results pointed to the possibility of silencing even after cell division [1]. Moreover, genes implicated in molecular memories may switch status surprisingly late in development or switch dynamically and have regulatory patterns that are far more complex than a single transition between on or off states [1]. Thus, there was a trend towards stable effects of epigenetic states on development, but with significant variability.

Research on the reversibility of epigenetic states has since moved beyond fruit flies, and spans multiple types of in vitro models, model organisms, and work on human tissues, in situations of both health and disease. Key areas of research include the determination of cellular identity during embryological development, modelled using induced pluripotent stem cells (iPSCs). iPSCs rely on a method whereby differentiated cells – such as a fully developed skin cell – can be reprogrammed to an undetermined state and then redirected to a new developmental path. Part of the enthusiasm for these cells comes from the fact that reprogramming to the undifferentiated state does not implicate any manipulation of the genome but relies on triggering epigenetic plasticity at regulatory elements implicated in cellular identity. In other words, interventions targeting the epigenome [5] may potentially rewrite cell fates by erasing or reversing memories of their pasts to produce cells perfectly identical to ‘true’ stem cells, which would amount to a process of full reversibility. However, it is now clear that iPSCs retain epigenetic traces of their previous differentiated state [6], suggesting an only partial reversal. Therefore, what scientists have referred to as silence (i.e. the return to undifferentiation), in these experiments, is only partially restored. This molecular plasticity underlying cellular identity over the cell lifespan argues against a binary model (e.g. with an epigenetic landscape as either mature or immature) and instead supports a gradual, context-dependent balance between persistence and reversibility. This emerging research echoes findings discussed by Ringrose and Paro regarding highly dynamic shifts or epigenetic traces of a cell’s history that resist experimental erasure.

Researchers working at the scale of the human lifespan do not necessarily depict such nuanced portraits of the dynamics of epigenetic states. Instead, they have argued that durable epigenetic states result from traumatic events [3]. (See also Keaney et al. in this volume, Chapter 14.) These epigenetic states are described as setting off brain alterations that contribute to psychological traits – such as impulsivity, interpersonal difficulties, or emotional lability – that ultimately potentiate the risk of mental illness. Research on reversibility – on a variety of species and scales – provides critical insights into human lifespan and DOHaD researchers. A careful review of this work reveals the considerable uncertainty about the dynamics of these processes. Yet it is only through a fine-grained understanding of such processes that scientists may conceive of how reversibility of epigenetic states may occur and how therapeutic interventions might silence molecular traces of past adverse events.

26.3 The Epigenetics of Memory Formation and Its Effects

Recent neuroepigenetics research advances that traumatic experiences may increase the risk of psychopathology through acquired molecular states etched into memories. The use of the term ‘memory’ in neuroscience research is polysemic, referring to a range of processes at different scales. In particular, it is often evoked in ways that are consistent with its common sense description, which roughly overlaps with the concept of episodic memory. Episodic memory, formally, refers to the ability to encode one’s life events and includes a range of cognitive functions that rely on interacting brain structures. The term molecular memory, by contrast, refers to molecular mechanisms correlated with any event leading to lasting cellular changes, whatever their implication in episodic memory, or any other brain property.

In this chapter, we are interested in epigenetic states as they are thought to correlate with the experience of past adversity (regardless of whether they may affect episodic memory or other physiological systems, for example reactivity to stress), and how they are believed to maintain – or not – a molecular modulation of gene activity: in effect, producing at-risk states. In order to determine which notions of memory are implicit or explicit in researchers’ hypotheses about whether molecular memories and their effects might be silenced, it is necessary to examine the uses of the concept of memory in epigenetics research.

A subset of researchers interested in memory and epigenetics have explored the so-called ‘“epigenetic code” in the central nervous system that mediates synaptic plasticity, learning, and memory’ [7]. In their models, neuroscientists Jeremy Day and David Sweatt evoke ‘the controversial theory of the “engram” – a (hypothetical) biophysical change in the brain that accounts for the material existence of memory (Josselyn et al., 2015: 201) … [and] suggest that epigenetic mechanisms, such as DNA methylation, may be a window into the brain’s memory’ [8]. They and other researchers became interested in how memory can be traced through epigenetic mechanisms in the brain, at a molecular level. Drawing mostly on research on model organisms, Day and Sweatt further argue that:

An interesting new understanding has emerged: developmental regulation of cell division and cell terminal differentiation involve many of the same molecular signalling cascades that are employed in learning and memory storage. Therefore, cellular development and cognitive memory processes are not just analogous but homologous at the molecular level. [7]

Their research presents cellular epigenetic and developmental mechanisms, and cognitive memory processes, as intertwined, and thus potentially actionable on a molecular level. In this understanding of molecular memory, the epigenome is ‘a crucial ‘missing link’ between life experiences and gene expression, which in turn will influence the ways in which neuronal circuitry and brain structures develop’ [8].

In these models, two characteristics of epigenetics are advanced, both of which we suggest should be approached with caution. First, that molecular memory may be homologous to episodic memory, and second, that epigenetics makes an exceptional contribution to the chain of events leading from life experience to the molecular memories of these events and their subsequent effects. Any proposition to silence memories of traumatic events would hinge on these relationships and the possibility of an intervention acting specifically on them. While Day and Sweatt put these ideas forward most explicitly, they implicitly inform many other researchers’ models of epigenetic memory and its potential reversal [9].

Day and Sweatt’s first proposition may be particularly misleading. Recent research indicates that every physiological function of the nervous system – such as feeding, sleep, or nociception – may implicate molecular mechanisms occurring in part through gene expression changes, under epigenetic regulation [10, 11]. In these studies, the role of molecular and epigenetic processes in the emergence and long-term regulation of those states appears similar to what has been identified in relation to the physiological function of episodic memory. This complicates any assertions of specificity or homology (besides the use of a common word) in the relationship between epigenetic and episodic memories. Episodic memory, instead, might be seen as affected by gene expression changes and epigenetic plasticity, much as the aforementioned other physiological functions, without necessarily being homologous to them.

The second proposition is similarly debatable. Responses to life experiences are complex and multi-scalar. In the case of trauma, their perception and encoding start with sensory processing of, for instance, sounds or movements, which are then cognitively apprehended by devoted brain areas, triggering negative emotions. Each of these operations relies on specialised cellular processes. At the sensory level, they include chemical (e.g. release of neurotransmitters in activated brain regions), physical (e.g. light sensing in the retina), or mechanical (e.g. transduction of sound waves by the tympanum) properties that act on temporal and spatial scales not necessarily compatible with or dependent upon any epigenetic plasticity. Moreover, it is the overall psychological impact of adversity, downstream of these multi-scalar processes, that is considered to trigger epigenetic changes.

Neuroepigenetic mechanisms are nonetheless widely considered to be implicated, to some extent, in the formation of molecular memories. Most of this research investigates DNA methylation, which we will focus on below. Changes in DNA methylation are considered not only to reflect past experiences but also to contribute to behavioural changes through, for example, the modulation of neuronal processes, heightened sensitivity to stress, and increased psychopathological risk. In terms of experimental designs, research on these processes is grounded in the triangulation of incongruent experimental designs. On the one hand, animal studies document how embodied epigenetic memories of early adversity may manifest in adulthood in controlled settings that limit confounding factors. Even in these studies, causal attribution of abnormal behaviour to epigenetic changes would require dedicated experiments that manipulate the proposed epigenetic substrate to prevent or reverse the abnormal behaviour (see next section). In humans, on the other hand, associations between adversity, epigenetic alterations, and later psychopathology are even more questionable. Sources of unaccounted variability over the lifespan, following trauma, are incomparably higher as studies typically analyse post-mortem brains of people who often die decades after experiencing adversity. Alternatively, peripheral ‘liquid’ biopsies (blood and saliva) that can be taken throughout life are more accessible but are less relevant for understandings of brain epigenetics. Thus, there is only a tenuous, associative relationship in animal and human studies between early adversity, epigenetic memories of these experiences, and drivers of later behaviours.

Ultimately, based on existing evidence, any delayed or long-lasting embodied memories are likely associated with multi-scalar adaptations, which include but are not exclusively encoded by epigenetic changes. Therefore, while epigenetic processes are plausibly recruited over the lifespan, during early adversity, and later when a host of related biological consequences mediate the impact of more recent life events, they do not operate in isolation. In this context, influential conceptualisations of epigenetic processes as exceptional contributors to molecular memories of past experiences appear to reflect an inability to place them in these long chains of back-and-forth, across temporal and spatial biological scales [12]. These limitations suggest caution when postulating relationships between life experiences, epigenetic modifications, and memory, particularly in the context of human adversity and psychopathology. Moreover, current understandings of this relationship might encourage attentiveness to the ways in which slippage between different types of memory explicitly or implicitly populates research on epigenetic plasticity and its potential silencing. This slippage contributes to conclusions that too easily conflate behavioural and molecular risk or resilience.

26.4 Experiments in Reversibility

In addition to efforts to understand the molecular mechanisms that may be associated with the experience of trauma and subsequent psychopathology, researchers are attempting to identify interventions that might reverse or modify epigenetic states and the psychopathology correlated with them. The most targeted epigenetic editing interventions aspire to modify the fundamental molecular processes associated with past experiences of trauma. Researchers hope that these modifications will affect neurobiological processes and, as a consequence, behavioural traits and reactivity to stress (e.g. as in the case of PTSD). The primary target of these interventions is not considered to be the factual or emotional content of an episodic memory such as the emotional relationship between the person and a specific object/event, but rather an affective state thought to be related to behaviour associated with past experiences of trauma. Affective states, in this perspective, are conceived as triggered neurobiological dispositions ‘operating outside the domain of consciousness and intentional action’ [13]. In this conceptualisation of neuropsychiatric risk, triggers are considered both devoid of exceptional qualities and sufficient to set into motion pathological responses. At their extreme, in certain neuroepigenetic research agendas, affective responses to triggers are thought to be sufficient to lead to suicidal acts [14].

Some of the research on the reversal or modification of epigenetic states focuses on well-established interventions such as antidepressants and psychotherapy. These therapies seek to mitigate the effects of past traumas through the alleviation of symptoms in the present (e.g. anxiety) and are now also studied for their effects on epigenetic mechanisms. This involves a reconceptualisation of these interventions as modulating basic affective states underlying clinically measured symptoms. Concerning antidepressants, researchers have associated several different epigenetic modifications (in the aforementioned peripheral samples, not the brain) with a positive response to antidepressants and are attempting to identify which epigenetic states might be able to predict responsiveness to these medications [15]. In a similar line of reasoning, researchers have suggested that epigenetic mechanisms may constitute ‘dynamic biological correlates of [psychotherapeutic] interventions’ [16]. However, the processes, directionality, or interactions linking symptom alleviation, intervention, and epigenetic states are far from comprehensively understood. For example, such research does not demonstrate whether (1) it is the intervention that reduces a person’s symptoms and this reduction subsequently impacts epigenetic profiles, (2) the intervention directly influences epigenetic plasticity, thereby modifying symptoms, or (3) some combination of the two. Therefore, at present, reasoning about the reversibility of behavioural and molecular states – and how they might relate to states of risk or resilience – remains muddled. This raises important questions about the inference of causality, as distinguishing between these possibilities would require direct and specific manipulation, or ‘editing’, of the epigenome.

Experimental approaches are being developed in rodent models to address the challenge of causal inference. Researchers such as Elizabeth Heller and Eric Nestler are attempting to carry out locus-specific epigenetic editing (i.e. affecting only a specific location in the genome [17]). Using this method, Heller and collaborators epigenetically reprogrammed a gene in a specific brain region to modify behavioural responses to later stress exposure, promoting susceptibility, or alternatively resilience, to this experience. They argue that the specificity of their approach allows them to understand how locus-specific epigenetic states may be causally implicated in the modulation of stress responses. The extent to which such manipulations are truly specific – affecting the targeted gene only – is unclear, with difficult technical and experimental challenges ahead. Nonetheless, these findings demonstrate the potential feasibility of intervening in targeted ways on the molecular processes implicated in stress or trauma responses to potentially silence molecular memories of past experiences.

Other researchers are drawing on different approaches to target the molecular machinery that may mediate epigenetic reprogramming. A team led by Moshe Szyf and Gal Yadid recently investigated a rat model of post-traumatic stress disorder (PTSD), in which they identified changes in DNA methylation [18]. In an attempt to undo PTSD-like behaviours, they manipulated the expression of one of the two enzymes responsible for methylating DNA in the mammalian brain (Dnmt3a). While the results offer support for the hypothesis that DNA methylation changes may contribute to PTSD-like behaviours, the evidence of causality through epigenetic reversibility may be considered more indirect than in the previous study by Heller et al. For instance, they did not identify if or how their manipulation of the enzyme directly affected the DNA methylation states that were triggered in the model, but instead reasoned by inference that the enzyme must have affected them. Despite these limitations, Yadid et al. suggest that it may be possible to translate their intervention to humans using a systemic therapy rather than direct manipulation in the brain [18]. They propose the addition of a chemical donor for methyl groups to our diets, which raises further questions about the specificity of the intervention. Indeed, systemic therapy would likely affect every cell in the whole body in which methylation of DNA affects their activities. Such an induction of epigenetic plasticity may have broad and potentially detrimental effects throughout the body.

Together, these approaches bypass existing symptom-oriented therapeutic interventions that are aimed at alleviating the emotional impact of distressing and presumably durable memories, and instead aim to directly reverse the molecular imprints of traumatic memories. In theory, they are more akin to an intervention targeting the aetiology of post-traumatic states, returning a person to the affective silence of an epigenetic landscape unmarred by (mal)adaptive shifts brought on by adversity. Such interventions would hypothetically target a range of regulatory elements. Systemic global methyl donor treatments, for instance, may have the potential – to use an analogy – to reopen critical windows of neuroplasticity among people who are biologically beyond the developmental period associated with early-life plasticity, when the effects of negative experiences are considered to be particularly harmful (Reh et al. 2020). In other words, the treatment is conceived of as affecting the canalisation that presumably takes place in a person’s life and sets them on a particular life trajectory [12].

It should be underscored that any judgement of a return to silence in this research might be considered arbitrary. At the extreme end of wiping cellular memories clean, as in the case of iPSCs, even efforts to epigenetically reprogramme cells back to stem cell states are unable to completely remove molecular traces of their past differentiated identity. In addition, it is clear that epigenetics is only one part of multi-scalar responses to life experiences, and whether the latter would be able to return to silence upon epigenetic editing is unknown. In terms of particular interventions, systemic therapies that aspire to modulate epigenetic processes come with the potential for sweeping effects on our bodily processes. Even targeted epigenetic editing interventions may either miss their mark (being unable to remove the molecular memories associated with past trauma) or destabilise people’s affective identities in unforeseen ways. The limitations of these interventions place the appraisal of a return to silence in a relative framework.

In addition, epigenetics research on the effects of trauma is grounded in the comparison of model organisms that were exposed or not. Yet the animals are not tested prior to exposure and interventions, an assessment that would be necessary to provide a glimpse of ‘before’, which would hypothetically reflect a state of silence. In humans, these before states are not tested either, given that brain tissue can only be studied post-mortem. Further complicating judgements of before, after, or a return to a previous unaffected state, research on inter- and transgenerational effects of trauma and long-term evolutionary inheritance of epigenetic states raises additional questions of whether before should be considered a state during an individual’s life or whether it should include in utero or preconception experiences, including potentially those of parents, or even longer time scales [19]. Thus, the before state to which a person would return is rarely assessed in this research, and questions remain as to when before should be identified in a person’s or a lineage’s trajectory. In the light of current understanding of molecular memories, it might one day be possible to reverse a single epigenetic state with a richer understanding of the processes involved, but we are not there yet. These understandings would necessarily include the kinetics, particularities, and potential reversibility of epigenetic processes in the brain; their reciprocal interactions with other levels of biological organisation; and, finally, the development of more precise interventions, targeting pathophysiological substrates only.

Limitations notwithstanding, the silence envisioned in these interventions for stress or PTSD spans ideas about memories and silencing through interventions targeting reversibility and plasticity, with epigenetic manipulations proposed as the key means of undoing the effects of past adversity. These perspectives integrate beliefs about the tendency towards stability of epigenetic states, as discussed by Ringrose and Paro over 20 years ago. They presume that molecular profiles are fixed and in need of molecular interventions to be righted. What is set aside, in terms of Ringrose and Paro’s analysis, is indeterminacy and the context and variability of epigenetic states: whether, and under what conditions, acquired epigenetic states may be reversed and with what effect.

26.5 Conclusion

The interventions described in this chapter aim to silence memories of the past to create an unencumbered present and future: by wiping a person’s past slate clean – whether in a targeted or more generalised way, depending on the intervention – it is assumed that the problem lies in the individual and only in their past as an isolated event/biomarker. Our goal has been to assess the state of knowledge about memories and their silencing and to consider the complexity of reasoning between molecular and experiential levels. We contend that researchers who aim to help people ‘remember silence’ should carefully reflect on the tenuous relationship between potential epigenetic and behavioural states of risk or resilience. We also argue for closer attention to the multi-scalar processes that may affect this relationship. Indeed, even if the trauma occurred in the past and a therapy was able to reverse an epigenetic state correlated with it at a later date, this does not mean that the multitude of multi-scalar processes associated with the traumatic event would also be silenced. Moreover, for many people who experience early-life adversity, ongoing trauma is as much an experience of the present as of the past [20, 21].

Ultimately, Ringrose and Paro’s essential provocation concerning the indeterminacy of epigenetic states remains a powerful reminder for research that frames epigenetic trajectories as linear and fixed. While we may seem closer to the possibility of remembering silence based on claims in emerging research about epigenetic reversibility, there remains a chasm between understandings of epigenetic reversibility and the emotional and affective states associated with what are considered states of neuropsychiatric risk or resilience.

Chapter 27 Disability in DOHaD and Epigenetics Towards Inclusive Practice

Kaleb Saulnier , Lara Azevedo , Neera Bhatia , Lillian Dipnall , Evie Kendal , Garth Stephenson , and Jeffrey M. Craig

27.1 Introduction

Developmental Origins of Health and Disease (DOHaD) and epigenetic research that investigate causal mechanisms and predictive biomarkers have often occurred in the absence of discussion of ethical, legal, and social implications or engagement with disability communities. This has often led to maternal blaming, labelling, stigmatisation, and ableism. Considering the debate on different models of disability by disability activists and social scientists, this is a timely opportunity to optimise the design of epigenetic research into conditions labelled as disabilities. Research aims should address the needs of disability communities, acknowledge diversity, and move away from medical to social models of disability.

Our chapter considers the implications of epigenetics research, as a mediator of DoHAD, for people with autism, an example of a condition some label a disability. We discuss how views on epigenetics and autism have changed over time, including how research can enhance the lived experience of autistic people through contributions to understanding how autism develops and how the strengths and needs of autistic people can best be identified and supported. We argue there is a need for researchers, including those with autism, to work with autistic people and their supporters to co-design studies promoting this understanding, centring autonomy and the provision of information to autistic individuals, including whether to engage with current and future epigenetic tests, particularly those available direct to consumers. In summary, we urge researchers planning such studies to first engage meaningfully and non-tokenistically with disability communities and continue to engage through to the writing and dissemination phases of their research.

27.1.1 On Terminology

Genetics research and autism studies have a complicated history, so we begin by establishing our choice of terminology and rationale for this. We acknowledge that there are strong and often polarising views about the issues presented in this chapter; however, we hope that we can contribute to meaningful discussion.

The principle ‘nothing about us without us’ communicates that decision-making that impacts a particular group should not take place without the full and direct participation of its members [1]. Thus, it is crucial that individuals be referred to using their preferred terminology. Person-first language evolved in the 1970s to separate the person from the descriptive trait, for example ‘a person with autism’ rather than ‘an autistic person’, and to give primacy to their identity as a person. Although this is well-intentioned, some disability activists have noted that this forced separation between person and trait reinforces the idea that disability is inherently negative and ignores the integral role that disability plays in shaping a person’s character and experience. As such, there has been a move towards identity-first language.

This terminology is by no means ubiquitous. A person-centred approach to language recommends that on an individual level, words that people use to self-describe should be prioritised.¹ For coherence, we have chosen to use ‘autistic person’ here, except where a direct quote incorporates other terminology. This is reflective of the preferred language identified by many autistic individuals and autism self-advocacy organisations [2].

27.2 Disability Politics in the Framing of Health

Disability studies emerged in the 1980s and engage with the concepts and consequences of disability, exploring, among other topics, what it means to be disabled in relation to the self and society [3]. Critical disability theory, which focuses on analysing and dismantling systems of ableist oppression, sits at the intersection of academia and activism [4]. Systems that privilege able-bodied people over those with disabilities are not only concerned with understanding the impacts of pathologisation but also undoing them. While there is no single approach to disability studies or politics, both consider the importance of centring and uplifting the stories, voices, and perspectives of disabled individuals in all disability work [1].

27.2.1 Models of Disability

Two dominant models of disability are often contrasted in literature and practice. The medical model ties disability directly to the body, focusing on possible interventions to bring it to a particular type of functioning [5]. The social model situates disability in the social context and physical environment of the individual and is focused on identifying barriers that prevent full participation in society. The latter model differentiates between impairments – attributes impacting how the body and brain operate – and disabilities – restrictions imposed by societal standards that reflect normative ideas of how bodies should function. A third model is ‘neurodiversity’, a term first coined by autistic sociologist Judy Singer and popularised by Steven Silberman in his book, Neurotribes, in which he defines it as follows:

the notion that conditions like autism, dyslexia, and attention-deficit/hyperactivity disorder (ADHD) should be regarded as naturally occurring cognitive variations with distinctive strengths that have contributed to the evolution of technology and culture rather than mere checklists of deficits and dysfunctions. [6]

Like the social model of disability, neurodiversity emphasises the disabling nature of stigmatisation and the prioritisation of brains classified as ‘normal’.

27.2.2 Disability in Research

To de-pathologise disability requires engagement with disability communities and scholars in developing frameworks from research design to knowledge translation. Sometimes referred to as participatory or community-engaged research [7], evidence indicates this approach contributes to better health and social outcomes. It is critical to respect the contributions of disabled scholars, activists, and organisations and promote collaboration between disabled and non-disabled researchers, and disabled participants and their advocates. Participatory research means an increasing understanding of disabled individuals as co-creators of scientific knowledge, rather than passive subjects.

There is, understandably, hesitation in disability communities regarding participation in medical research. As with many vulnerable communities, the history of unconsented research and other research harms is long and fraught [8]. Community members are quick to spot ableist rhetoric and stigmatisation in research documentation and are reluctant to participate if their bodies, lives, and experiences may be used to pursue goals not aligned with the expressed needs of disabled individuals. Thus, participatory research does not begin with inviting disabled individuals as research subjects, but rather, with listening, learning, humility, and trust-building on the part of non-disabled researchers, using the principle of co-design and through participant advisory groups.

27.3 Mapping Disability onto DOHaD and Epigenetics

As discourse shifts from the medical model, bioethicists and clinicians have begun to recognise how social factors play a primary role in the treatment of disabled individuals. The DOHaD model represents a particularly fruitful opportunity for this shift, focused on a bio-psycho-social model of health and disease [16]. Similarly, epigenetics’ attention to the role of environment, exposures, and stress moves away from the biological determinism of the genomics era [17] towards a more holistic understanding of health. Nonetheless, researchers in DOHaD and epigenetics should refrain from importing potentially harmful presumptions into these emerging fields, with bioethicists and disability scholars already expressing concerns that applying the medical model to these areas risks intensifying rhetoric around responsibility and blame for social and environmental exposures, particularly when associating maternal exposures with future disability [18, 19].

The DOHaD phenomenon is supported by ample animal and human evidence but has an intrinsic focus on ‘health vs disease’. This neglects natural variations not classified as ‘health’ or ‘disease’, including a wide range of ongoing or recurring behaviours, cognitions, and health conditions that are multidimensional. These include neurodiverse conditions such as autism, whose communities refute the labels of ‘disease’ and ‘disability’, similarly to the deaf community. This is relevant when attempting to apply epigenetic models of disabilities to traits that cannot reasonably be classified as ‘disease symptoms’. Therefore, we have a responsibility to be careful with terminology when engaging with participants from the disability community, including when planning and reporting epigenetics research.

27.4 The Case of Autism

Autism presents a valuable case study to explore the intersection of disability, DOHaD, and epigenetics, as a condition that has long oscillated in medical and public imagination between having social, environmental, or biological origins. DOHaD research has shown that early-life exposures to social, biological, and environmental factors can influence fetal development. Influential biological factors include maternal infection and inflammation, which can lead to a state of maternal immune activation where immune regulatory mediators are expressed in higher-than-normal ranges, a possible risk factor for autism and other neurodevelopmental and psychiatric conditions [20, 21]. Suboptimal nutrition before or during pregnancy, particularly vitamin B9 (folic acid), has also been implicated [22], as well as prenatal exposure to traffic-related air pollution and some insecticides [23]. Factors that cannot be explained by shared genetics and environment have also been associated with autistic traits, for example in one twin from a genetically identical pair.

Despite a strong genetic influence, there is considerable genetic heterogeneity across autistic individuals [24]. Around 5 per cent are also diagnosed with a clinically and genetically diagnosable syndrome, and around 15 per cent can be attributed to simple genetic changes such as single gene mutations or copy number variations. For the remaining individuals, evidence points to autism as a polygenic condition, that is resulting from genetic differences spread across hundreds, possibly thousands of genes [24]. These genes appear commonly involved in brain development, epigenetic regulation of gene activity, and metabolism, suggesting possible causal mechanisms for autism. Since the early 2020s, autism-associated variants have been grouped together to form a ‘polygenic risk score’, with a higher score theoretically indicating a higher likelihood of autism [25].

There are more genetic differences in genes encoding components of epigenetic mechanisms in autistic people as a group compared to non-autistic people. As their gene products are likely to act at multiple genomic regions, some autism-specific epigenetic differences will likely have strong genetic components, [26] increasing their likelihood of being stable over time and therefore useful as diagnostic and prognostic biomarkers. Associations between epigenetic states in the sperm of fathers of autistic children compared to those with neurotypical children [27] are more likely to be explained by genetic factors, unless genetics are controlled for, for example, in identical twin studies.

Epigenetic studies of autism have identified similar genes and gene functions to genetic studies, including those involved in epigenetic regulation and synaptic function. However, far more immune system genes have been identified in epigenetic studies of autism diagnoses [34]. Epigenetic studies have also investigated specific dimensions of autism, for example social communication [28], potentially predicting biomarkers at birth [29], and risk scores [30]. However, these findings have yet to be replicated.

27.4.1 The Social Construction of Autism

Criteria for autism in the Diagnostic and Statistical Manual (version 5) include but are not limited to ‘persistent deficits in each of three areas of social communication and interaction plus at least two of four types of restricted, repetitive behaviours’ [31]. This deficit model, at times focused on the external viewer’s perception of autistic experiences, typically shapes research seeking to minimise these behaviours and accompanied distress. By contrast, neurodiversity-focused groups frame autism as a constellation of strengths and challenges across social and sensory spectra and focus on research and resources to support autistic individuals in achieving their best quality of life [32].

Another current view considers autism as a potentially disabling condition that nevertheless may confer various positive traits [33]. However, in many cases this view has merely re-circumscribed capitalist values of productivity, for example celebrating those autistic traits, such as hyperfocus, that can be exploited by employers to improve work output. While this view has partly enhanced our understanding of neurodiversity and the need for better neuroergonomics in the workplace, the ultimate focus has not been on promoting quality of life for autistic people. Moreover, other autistic traits that are considered neutral or positive within the autism community, such as stimming to self-soothe and express emotions, are still misunderstood as negatives or viewed with discomfort.

The experiences of autistic individuals in healthcare provide an opportunity for examining pathologisation of their condition via scientific research into DOHaD and epigenetics. These research areas rely heavily on the interpretation of links between social and other environmental factors with biological outcomes. Perhaps the most widely recognised image of the autistic individual is that of a white, masculine-presenting child with an inability to make eye contact, limited or stilted speech, and a fascination with patterns or trains. This perception has recently begun to shift, prompted in part by an increase in later-in-life diagnoses in cisgender women as well as non-binary individuals and transgender men and women, who may present differently from this stereotype.²

Autism and diversity of gender identities and experiences overlap substantially, further impacting access to appropriate support and care. Gendered differences in presentation have led autistic girls and women to be underdiagnosed, misdiagnosed, or diagnosed at a later age, sometimes only after their own child’s diagnosis [34]. This discrepancy has contributed to a lack of understanding of key mental health conditions that co-occur alongside autism, including eating disorders, depression, anxiety, and suicidality. Similarly, disparities in the impact of race and ethnicity on the timing and frequency of diagnosis have led to a paucity of resources and support for racialised/ethnic minority autistic youth, who at the same time experience increased risks and rates of police and other state-sanctioned violence and incarceration [35]. There is an urgent need for an intersectional approach to all disability research, but particularly epigenetic studies examining the social and environmental contexts for the lived experiences of autistic and other disabled individuals.³

27.5 Reframing Epigenetics Research to Address the Needs of People with Disabilities

27.5.1 Biomarker Development for Conditions Classed as Disabilities

We are still far from having reliable predictive or diagnostic genetic or epigenetic biomarkers for conditions such as autism. One major factor that clouds the interpretation of such research is study design. Most classify autism as one entity, whereas it is a highly heterogeneous condition. Furthermore, co-occurring conditions such as ADHD are largely ignored in such studies. Some researchers have turned away from a categorical to a dimensional approach to the origins of autism, using continuously variable dimensions such as anxiety, attention, sensory processing, specific interests, repetition, social interaction, and communication [36]. We suggest that this method is preferred because it targets traits that can be clinically defined and can identify areas of strength as well as areas in which autistic individuals may require understanding and assistance. This approach also captures intersecting dimensions of co-occurrences, such as ADHD, for example, sustained attention.

We suggest that future studies be based on dimensions of autism with a view to meeting the self-determined needs of autistic individuals and the autism community. A dimensional approach also reflects the spectrum of neurodiversity within and outside the autism community and the reality of the social model of autism rather than the medical model.

As the field of epigenetics moves towards identifying more biomarkers for conditions and associating these with developmental, social, and environmental correlates, the rhetoric surrounding curative approaches to disability could increase. This rhetoric is closely tied to medical and deficit models of disability, with their foundational assumptions that people with disabilities wish to be rid of the disabled parts of themselves. For some, this may be true; the existence of the social and neurodiversity models does not detract from the struggles precipitated by certain features associated with disability, such as chronic pain, anxiety, loss of quality of life, or early mortality. Rather than attempting to categorise disabilities wholesale as ‘bad’ (e.g. where we may aim to repair or alter the body or brain) or ‘good’ (where we may instead target disabling factors in the society or environment), a more useful account would examine components of disability that are unwanted by the individual who experiences them. Again, following the principle of ‘nothing about us without us’, it is important to differentiate between calls for prevention and cure that come from researchers and healthcare providers, and policies based on the lived experiences of disabled individuals and their advocates. In doing so, a stark divide can appear between the expectation of the disabled experience and the reality.

It has long been argued that health economic metrics, such as the quality-adjusted life year (QALY) or disability-adjusted life year (DALY), are not sensitive to the real experiences of disabled people and ignore the significant adaptive ability of individuals [37]. Inviting more conscious consideration of disabled peoples’ own experiences of their disability can also avoid the tendency to objectify disabled persons’ bodies and view them as separate from the disabled experience. This helps avoid the risk of ignoring meaningful needs assessments conducted by the community. In other words, embracing a neurodiversity model does not mean neglecting to provide support for autistic individuals who consider certain traits to be personally disabling or undesirable. Ethically, it is important to remember another classic phrase in the disability community, coined by autism advocate Dr Stephen Shore: ‘If you’ve met one person with autism, you’ve met one person with autism’ [38]. Again, the diversity of manifestations and personal experiences can only be incorporated effectively into epigenetics and genetics research if studies are co-designed and guided by diverse members of the autism community.

27.5.2 Direct-to-Consumer (DTC) Epigenetic Tests

In the past five years, there has been a growing number of companies selling epigenetic tests directly to consumers, that is without the need for a referral from a healthcare practitioner [39]. Despite being unable to define a ‘healthy’ epigenome, DTC companies focus on identifying epigenetic biomarkers in consumers’ blood or saliva samples with the promise of enabling consumers to improve their health outcomes. An ‘altered’ epigenetic status could indicate early-life exposures that increase the likelihood of developing certain conditions, which could be targeted for intervention due to the potential reversibility of epigenetic changes. Identifying environmental risks for the development of a condition also means that prevention strategies could be adopted to reduce the chance of its development, for example, via diet and exercise changes. In the case of autism, there are currently tests being developed to facilitate diagnosis in children as young as 18 months old [40]. Here, the promise is to provide biological data to complement more subjective analyses to expedite autism diagnoses and access to early intervention and resources.

However, DTC epigenetic testing raises various ethico-legal issues, related to the core technical issue surrounding the precision of epigenetic biomarkers for diagnosing complex conditions. Marketing that overestimates the reliability of epigenetic test results could exploit consumer trust in science to sell a product that falls short of its promises. Test results could also affect an individual’s access to insurance policies, particularly life insurance, as the reporting of test results is often a legal obligation of the applicant. With a focus on environmental risks, there is also a tendency to blame individuals for the development of associated conditions. Here we acknowledge the long history of blaming mothers for autism, a fact well demonstrated by the term ‘refrigerator mothers’⁴ often used to describe them [19]. While parents of children with autism seem to support the development of epigenetic testing for improvement of the diagnosis process [41], there is a need for clear regulations of the DTC market to protect consumers, especially vulnerable populations.

27.5.3 Ethical, Legal, and Social Implications of DoHAD Research for People with Disabilities

From an ethical perspective ‘respect for persons’ is one of the fundamental tenets of Western biomedical ethics. Its application in DOHaD research is often more complex than in standard clinical care [42]. Core ethico-legal issues here include maintaining confidentiality and privacy, and gaining informed consent for medical interventions, which work together to promote autonomy. While protecting sensitive information such as medical diagnoses and treatment decisions may be relatively simple within the practitioner–client communication paradigm, genetic information, for example, might be problematic for the individualistic Western ethico-legal model. (See Karpin in this volume.) From a DOHaD perspective, genetic and epigenetic information might best be conceived of as family information, making privacy concerns more complex. However, the rationale behind protecting this information remains the same: promoting autonomy, including through avoiding potential coercion from those who would misuse sensitive information to discriminate against individuals. The latter is relevant for accessing employment, health and life insurance, and healthcare services. Whether DOHaD and epigenetics research should aim for family or community, rather than individual consent, falls beyond our scope here, but we recognise that when a test impacts more than the individual, there is potential for social harm against others who are impacted by the results. For research on communities with disabilities, especially those with a potential genetic contribution, this suggests the co-design of research studies is important to ensure knowledge about inheritance is not weaponised against the community or used to engage in blaming or labelling of parents or offspring.

Confidentiality is a key pillar in the doctor–patient relationship protected under common law and statutory regulation. For example, privacy laws in Australia governed under the Privacy Act 1988 (Cth) have a broad reach, protecting a range of information, including health information. According to the ‘For your information: Australian Privacy Law and Practice’ ALRC Report No 108) [43], ‘privacy’ covers several aspects, including data protection, such as medical and government records; bodily privacy, such as invasive procedures that may include genetic and epigenetic tests; and communication, such as emails.

The disclosure or privacy of sensitive genetic information in some instances might be problematic. For the disability community, it is possible that epigenetic data collected from one consenting individual or family may have immediate relevance to other community members. For this, the right ‘not to know’ might be as important as the right to know the genetic factors involved in the development of autism. Importantly, once these data exist, they may have wide-ranging impacts on members of the community who did not consent to the research.

27.5.4 Incorporating Perspectives from Multiple Stakeholders

A goal of disability activists has been to reframe conversations about disability, health, and disease away from views that centre concepts of ‘normalcy’ and ‘functionality’ and to instead centre the disabled individual as the core stakeholder in the discussion of their own body and experience. Throughout the twentieth century in particular, the concept of ‘wellness’ came to be equated with ‘virtue’, situating the body as a ‘site for moral action’ [44] with regard to the pursuit of health. The medical model, in addition to enforcing the idea of a ‘normal’ state of the body to which its owner should aspire [45], increasingly pushed a ‘functionality’ argument that privileged a body’s capacity to contribute to labour [46], and disdained disability precisely because of the implication that the disabled individual is of inherently reduced worth under capitalism.

The term ‘stakeholders’ is suggestive of a consumer-driven approach to health and well-being that places disabled individuals immediately at a disadvantage [47]. As a result, the stakeholders most often centred on disability research have been medical practitioners and families of disabled individuals. While both caregivers and practitioners have a significant interest in the disability conversation and valuable experiences to contribute, at times, this has come at the expense of the voices and narratives from disability communities and their advocates. In autism research, this has contributed to frustration and conflict. As this research moves forward, it must include autistic individuals and their caregivers where necessary (as participants, researchers, and scholars) at the centre of the conversation from research development to knowledge translation.

27.6 Conclusion: Recommendations for Engagement with Disability Communities in DOHaD and Epigenetic Studies

It is essential to engage with disability communities and their supporters at every step from research design to knowledge translation. Previous experiences within these communities highlight the risks that genetic research can lead to discrimination and stigmatisation, and in the case of DOHaD, this extends not only to individuals with the condition of interest but also to their parents [48]. For this reason, we advocate for more inclusive research practices that build trust with disability communities, listen to their needs, and promote support, while maximising autonomy, dignity, and respect for all members of the community.

In the case of autism, we call on researchers to reflect on their motivations when planning epigenetic studies of autism, considering whether predictive testing prior to the typical onset of symptoms would allow for early modes of support [49]. We urge researchers to seek advice from the autistic community when studying environmental contributions to autism to consider structural frames aimed at policy change in addition to those focused on the agency of individuals. Researchers should also be mindful of the language they use in planning and reporting research findings and of adopting a dimensional framework for cognitive assessment.

Future studies may look at the ethical implications of handling and releasing wide-scale epigenetics research data on autistic communities to ensure knowledge is used to meet the needs of this community and improve the quality of life. In summary, we urge researchers planning DOHaD and epigenetics research to listen to and engage with disability communities when they say, ‘nothing about us without us’.

Chapter 28 Creating Good Data Our Way An Indigenous Lens for Epidemiology and Intergenerational Health

Sarah Bourke and Raymond Lovett

28.1 Introduction

Momentum is building for epidemiological research led by and for Indigenous peoples. Backed by a human rights agenda, this drive is gaining speed due to wider calls to decolonise the social sciences, and increased recognition of the critical importance of incorporating Indigenous expertise in the conceptualisation, development, and execution of effective health research. ‘Decolonising’ epidemiology may involve several processes that are best determined by the communities who contribute their data. In Australia, this has involved acknowledging the complicity of the sciences in the colonisation of Aboriginal and Torres Strait Islander lands, waters, skies, and peoples, and the need for data that reflect the ongoing impact of settler-colonial practices and ideals on our communities. This work has often been conducted within a strengths-based framework, emphasising the inherent assets and resilience of Aboriginal and Torres Strait Islander communities, and the role of culture as the foundation of our individual, social, ecological, and spiritual health and well-being. To decolonise epidemiological research about Indigenous communities, Indigenous peoples must be in control of the definition, collection, and use of their data, following Indigenous Data Sovereignty (IDS) and Indigenous Data Governance (IDG) protocols, which ensure these data serve our self-determined interests and futures.

Hoke and McDade argue that lifecourse interventions based on DOHaD research are ‘rarely situated within the cultural, social, or political-economic context of the populations examined’ [1, p. 190]. Further, there is a lack of research focus on intergenerational or transgenerational events impacting long-term adult health in subsequent generations:

Rather than acknowledging the maternal body as the product of ongoing physiological, social, and political-economic processes, these influences on maternal physiology are often placed within an analytical black box and ignored. [1, p. 191]

Having a better conceptual understanding of the contexts in which people live would allow DOHaD researchers to design better measures to capture data on those concepts and then monitor any changes when interventions are put into place. We suggest that epidemiological methods that centre Indigenous lifeworlds have the potential to measure contextual influences and to identify salutogenic and protective factors that support holistic health and well-being for Indigenous peoples.

This chapter provides an overview of current discourses around centring Indigenous ontologies (ways of being), epistemologies (ways of knowing), and axiologies (ways of doing), also known as Indigenous ‘lifeworlds’, in epidemiology with a particular focus on Indigenous Australian (Aboriginal and Torres Strait Islander) perspectives. Mayi Kuwayu, the National Study of Aboriginal and Torres Strait Islander Wellbeing, will be used as a key example illustrating how epidemiological research may be led, owned, and governed by Indigenous peoples to produce rigorous and meaningful data that reflect Indigenous lifeworlds. Centring Indigenous perspectives may provide valuable tools for the development of the DOHaD lifecourse framework and future studies that seek to address holistic determinants of intergenerational health and well-being.

28.2 Part 1. Centring Indigenous Perspectives in Epidemiology

28.2.1 The Colonial Project as an Origin of Ill health and Disease

It is well known that colonisation and colonialism are determinants of Indigenous ill health [2]. Colonisation by nations including Britain, France, Spain, and others resulted in the mass genocide and/or displacement of Indigenous peoples across the world, and the methodical dismantling of Indigenous lifeways, undermining millennia-long connections to land, culture, and kin. Colonialism, as enacted through historical and ongoing contemporary colonial processes, encompasses a range of risk factors for health and well-being [3]. This includes interpersonal, institutional, and political processes that, at a minimum, aim to control or dominate a population and, at the extreme end of the spectrum, aim to eliminate or exterminate the population.

In Australia, it is estimated that around 90 per cent of the Aboriginal and Torres Strait Islander population lost their lives between 1788 and 1901 because of widespread colonial violence, introduced diseases, and the theft of land and water resources by British settlers [4, 5]. Those who survived were subjected to systematically racist and discriminatory programmes and policies. Many were dispossessed of their traditional lands and forced to live on Christian-run missions, which often forbade the use of any language other than English and actively suppressed important cultural practices. One particular assimilation policy enacted from 1910 to 1970 in Australia led to 11–24 per cent of all Aboriginal children being stolen from their families by government agents and held in Christian-run institutions for ‘re-education’ and training as domestic servants for the White middle classes [6]. They are known collectively as the Stolen Generations. Colonisation has thus generated significant intergenerational trauma for Indigenous peoples, reinforced by ongoing and cumulative ‘biosocial injury’ inflicted by settler-colonial nation states and individuals [7].

Settler-colonialism imposes cultural values, religions, laws, and policies that often go against the rights and values of Indigenous peoples. The concept of settler-colonialism as a health risk factor is well understood within Indigenous populations [8–10]. However, in Australia there is limited ability to examine such links due to an absence of epidemiological data on settler-colonial exposures. This itself is a manifestation of racism, given the lack of research priority and attention to settler-colonialism and its impacts. White racism and discrimination against Indigenous peoples thus perpetuate colonial trauma.

A small number of studies demonstrate poorer outcomes for Aboriginal and Torres Strait Islander peoples who experience individual settler-colonial factors such as discrimination [11–15] and the Stolen Generations [16] versus those who do not. However, most, if not all, of these studies are small-scale, use measures that have not been validated, and are not population representative. In addition, there are no quantitative studies that identify factors that can act as a buffer against the negative impacts of settler-colonialism. This is critical in taking the next step past identifying associations and towards identifying practical, strengths-based solutions to guide policy and programme development.

Research conducted by Indigenous peoples, for Indigenous peoples, and with Indigenous peoples is a direct challenge to the ongoing colonisation of our lands, cultures, and communities. As Tuck and Yang [17] stated, ‘Decolonization is not a metaphor’. It unsettles the settlers. Indigenous peoples have always conducted research, and we have always used counting as a tool for this research. This fact alone frequently unsettles non-Indigenous researchers in the field of epidemiology who often state that issues of importance to Indigenous peoples, such as culture (the practising of it or revitalisation of it), cannot be quantified. The centring of Indigenous peoples in epidemiology therefore means accounting for settler-colonial-inflicted biosocial injury and centring Indigenous definitions of health and well-being and the determinants thereof within research.

28.2.2 Indigenous Definitions of Health and Well-being

Indigenous identities and concepts of health and well-being are fundamentally connected to the land within an eco-centric relationality or kinship system that defines social and spiritual obligations to family, community, and the land [18, p. 200]. This relationality is reinforced through cultural frameworks where relationships to the past, ancestors, land, and the present are articulated through language, song, dance, storytelling, and maintaining traditional homelands, beliefs, and kinship [19]. As a result, Indigenous definitions of health and well-being are inherently holistic. The Medicine Wheel, or Sacred Hoop, has been used across Turtle Island (North America) to convey Indigenous philosophies of well-being using four quadrants (sometimes represented by the four cardinal directions – North, South, East, and West) to represent the connectedness between the physical, emotional, mental, and spiritual elements of life and the need for balance across these four areas. In Aotearoa (New Zealand), Māori holistic health and well-being also relies on a balance across four fundamentally interconnected elements – wairua (spiritual), whānau (extended family network), hinengaro (the mind), and tinana (physical) [20, p. 1141].

In the Australian context, the most used definition of Aboriginal and Torres Strait Islander health is as follows:

‘Aboriginal health’ means not just the physical wellbeing of an individual but refers to the social, emotional and cultural wellbeing of the whole Community in which each individual is able to achieve their full potential as a human being, thereby bringing about the total wellbeing of their Community. It is a whole-of-life view and includes the cyclical concept of life-death-life.… Health to Aboriginal peoples is a matter of determining all aspects of their life, including control over their physical environment, of dignity, of community self-esteem, and of justice. It is not merely a matter of the provision of doctors, hospitals, medicines, or the absence of disease and incapacity. [21, pp. ix–x]

In recent years, several studies have focused on expanding this definition. Garvey and colleagues [22] outlined their Fabric of Aboriginal and Torres Strait Islander Wellbeing model based on qualitative research with 359 participants, weaving together eight aspects of well-being: culture; community; family; belonging and connection; holistic health; purpose and control; dignity and respect; and basic needs. Butler et al. [23] conducted a comprehensive national literature review to identify nine interconnected domains of Aboriginal and Torres Strait Islander well-being, including autonomy, empowerment, and recognition; family and community; culture, spirituality, and identity; Country; basic needs; work, roles, and responsibilities; education; physical health; and mental health. Salmon and her colleagues [19] undertook an international literature review to identify and describe six key cultural domains essential for Aboriginal and Torres Strait Islander well-being: connection to Country; Indigenous beliefs and knowledge; Indigenous language; family, kinship, and community; cultural expression and continuity; and self-determination and leadership. Maintaining and reviving connections to land and culture has been found to be protective of Indigenous health and well-being across a range of studies in Australia and internationally [19, 24].

28.2.3 Indigenous Rights to Data

Indigenous definitions of well-being and its determinants have not been reflected in large-scale epidemiological data collections. For decades, statistics about Aboriginal and Torres Strait Islander communities have been wholly based on the perspective of the White settler-colonial state [25]. Hundreds of Aboriginal and Torres Strait Islander cultural groups were homogenised under the umbrella term ‘Indigenous’ for comparison to the (equally homogenised) ‘non-Indigenous’ population. This comparison was based on White settler-colonial definitions of health, social, and economic achievement. Sociologist Maggie Walter (Palawa) argued that the result of this comparison was the production of ‘5D Data’, emphasising the Difference, Disparity, Disadvantage, Dysfunction, and Deprivation experienced within the Aboriginal and Torres Strait Islander population [26]. She writes:

Current Australian practices in regard to the collection of data on Indigenous people are the cloned descendants of the data imperatives of colonisation. In what I refer to as the deficit data/problematic people (DD/PP) correlation, processes of enumeration have long been used to correlate the highly observable societal Aboriginal and Torres Strait Islander inequality with the concept of racial unfitness. [26]

When data stressing the ‘Indigenous problem’ present in Australian society are used to inform public health and policy development, the inevitable outcome is a re-colonisation of Aboriginal and Torres Strait Islander communities through programmes and policies designed to correct these perceived deficits. Interventions built on this deficit premise are prone to failure, as has been emphasised time and again by the Australian Government’s own monitoring scheme called Closing the Gap, established in 2008. The ‘Gap’ refers to the statistical gaps highlighted by the above-mentioned data, particularly the difference in life expectancy at birth between the Indigenous and non-Indigenous population, which currently sits at 7.8 years for females and 8.6 years for males [27]. According to Government statistics, over half (53 per cent) of the health ‘gap’ between Aboriginal and Torres Strait Islander Australians and the rest of the Australian population is accounted for by just five social determinants (employment and hours worked, highest non-school qualification, level of schooling completed, housing adequacy, and household income) and six ‘health risk factors’ (binge drinking, high blood pressure, overweight and obesity status, inadequate fruit and vegetable consumption, insufficient physical exercise, and smoking) [28]. The remaining 47 per cent of the health gap is currently unaccounted for.

Research seeking to measure the health of Indigenous peoples but exclude them from the development of such research is, quite simply, bad science. ‘Bad’ in this case may be interpreted in two ways. First, in the production of inaccurate, under-powered, and misleading data that, in Australia and Aotearoa at least, have informed decades of largely ineffective public health policy [29, 30]. Indigenous data must have equivalent explanatory power to non-Indigenous data to achieve health equity [31]. Second, it is now widely considered to be unethical to exclude Indigenous peoples from the development of research about their communities. The United Nations Declaration on the Rights of Indigenous Peoples (UNDRIP), which enshrines the right to self-determination for Indigenous peoples [32, Article 3], helped to spur a change in the way Australian human research ethics committees considered applications. Many major research ethics and funding bodies now require applicants to actively partner with Aboriginal and Torres Strait Islander individuals, organisations, or communities involved in the proposed research [33].

To counter the deficit discourse of Indigenous health that has been informed by the 5D data approach, Indigenous Data Sovereignty (IDS) and Indigenous Data Governance (IDG) protocols and principles have been developed for research based on the UNDRIP. For First Nations communities in Canada, this has been solidified through Ownership, Control, Access, and Possession (OCAP®), which became a registered trademark of the First Nations Information Governance Centre (FNIGC) in 2015 [34]. In Aotearoa, the Kaupapa Māori approach has been practised for decades, which upholds Māori self-determination and ways of knowing, doing, and being in research communities. In Australia, possible protocols and principles were discussed at a meeting in 2018 with delegates from the Maiam nayri Wingara Indigenous Data Sovereignty Collective and the Australian Indigenous Governance Institute. They defined Indigenous data as ‘information or knowledge, in any format or medium, which is about and may affect Indigenous peoples both collectively and individually’ [35]. IDS and IDG were defined as follows:

‘Indigenous Data Sovereignty’ refers to the right of Indigenous people to exercise ownership over Indigenous Data. Ownership of data can be expressed through the creation, collection, access, analysis, interpretation, management, dissemination, and reuse of Indigenous Data.

‘Indigenous Data Governance’ refers to the right of Indigenous peoples to autonomously decide what, how and why Indigenous Data are collected, accessed, and used. It ensures that data on or about Indigenous peoples reflects our priorities, values, cultures, worldviews, and diversity.

In line with these definitions, they determined that Indigenous peoples in Australia have the following rights [35]:

• To exercise control of the data ecosystem including creation, development, stewardship, analysis, dissemination, and infrastructure.

• To have data that are contextual and disaggregated (available and accessible at individual, community, and First Nations levels).

• To have data that are relevant and empower sustainable self-determination and effective self-governance.

• To have data structures that are accountable to Indigenous peoples and First Nations.

• To have data that are protective and respect our individual and collective interests.

Supporting these rights generates ‘good data’, which is the antithesis of 5D Data. The concept of good data extends existing global conversations around ethical data and data justice and incorporates IDS and IDG principles to describe a resource that Indigenous communities may use to address their self-determined interests and needs [36].

It is only since 2020 that Aboriginal and Torres Strait Islander representative organisations have been meaningfully engaged by Australian governments in the development of the Closing the Gap agenda and goals. Part of this engagement involved the recognition that Aboriginal and Torres Strait Islander peoples have the right to define our own needs and priorities in the policy arena. This partnership was formalised through the National Agreement on Closing the Gap (the Agreement) that commits to a genuine partnership between all Australian Governments and the Coalition of Aboriginal and Torres Strait Islander Peak Organisations [38]. The goal of this partnership will be to improve the life outcomes of Aboriginal and Torres Strait Islander people, acknowledging that supporting and strengthening Aboriginal and Torres Strait Islander cultures is necessary to achieve this goal [37, p. 4]. Large-scale epidemiological data that reflect Indigenous lifeworlds and reinforce IDS and IDG will be key to driving this national agenda.

28.3 Part 2. Mayi Kuwayu and the Future of the DOHaD

28.3.1 Overview of Mayi Kuwayu

Cohort studies, a type of analytical study, have made a considerable contribution to our understanding of human health and have been at the forefront of identifying the influence of social, environmental, and biological processes on health and well-being outcomes. The goal of analytic studies is to identify and evaluate the causes or risk factors of diseases or health-related events [38]. Cohort studies often involve identifying a ‘group’ of people to study and plan the research in advance, collecting data over time. Epidemiology, as the study of the patterns and distribution of disease and arguably health, is often the mechanism or method used in telling the story about the presence or absence of disease and what the likely relationships are between social, environmental, and other exposures under study. Due to their long time frames, expense, and difficulties in proving causation, cohort studies are uncommon when compared to other study designs.

Indigenous cohort studies are not common in settler-colonial states such as Australia, Aotearoa, and Turtle Island (referred to collectively as CANZUS) where Indigenous peoples are often incidentally recruited into studies. This limits the ability to conduct robust analytical studies specific to Indigenous groups, and, critically, the variation in Indigenous populations within CANZUS countries presents numerous challenges. The result is a ‘tyranny of the majority’ and evidence production that is biased towards the majority population while being underpowered for minority groups within those same cohorts. Additionally, these cohorts do not include risk exposures unique to minority groups, nor do they include exposure to unique potentially protective factors.

When the practice of epidemiology is led by Indigenous peoples its potential to contribute towards the achievement of health equity increases exponentially. Mayi Kuwayu is Australia’s largest longitudinal cohort study of Aboriginal and Torres Strait Islander well-being. As of January 2024, over 12,000 individuals have participated in the study that surveys Aboriginal and Torres Strait Islander-identified social and cultural determinants of health and well-being [39]. The Mayi Kuwayu baseline questionnaire was developed in consultation with Aboriginal and Torres Strait Islander community members, organisations, and research experts over the course of four years from 2014 to 2018 and is revised every two to three years. It includes a range of metrics that reflect community-identified determinants of health and well-being, such as connection to Country (e.g. ‘How much of your life have you lived on your tribe’s (mob’s) Country or Island?’) and cultural knowledge and practice (e.g. ‘How much time do you spend learning culture, kinship and respect?’). Mayi Kuwayu data adhere to IDS and IDG principles, and all requests for access to and use of the data must be approved by the Mayi Kuwayu Data Governance Committee, which is a group of Aboriginal and Torres Strait Islander community members and external researchers.

The intent is to conduct world-first analytical work to provide a robust understanding of how settler-colonial risk factors undermine the health and well-being of Aboriginal and Torres Strait Islander peoples, and how culture, a core strength of Aboriginal and Torres Strait Islander communities, can mitigate these adverse effects. This future work aims to understand some of the unexplained 47 per cent health inequity currently experienced by these communities in Australia, while also identifying how the very essence of Indigeneity (cultural maintenance, strengthening, and expression) is fundamental to improving health and well-being and reducing inequities [40, 41]. This work is critical in that settler-colonialism as manifested in historical and contemporary trauma is likely to have a profound impact on adult and intergenerational health.

28.3.2 Mayi Kuwayu Data and Relationship to DOHaD

The Mayi Kuwayu study was developed in response to calls from Aboriginal and Torres Strait Islander peoples to ensure health and well-being concepts were appropriately measured and captured. It is underpinned by a social epidemiological framework, concerned with the influences of social structures, institutions, and relationships on health and well-being [42, 43]. Therefore, the study is designed to enable the examination of health and well-being, taking into account the varied contexts in which Aboriginal and Torres Strait Islander peoples live, including diversity in exposure to settler-colonial factors, and diversity in opportunities to engage in cultural practice and expression. Further, the study is ideally placed to explore and quantify if, and to what extent, culture buffers the impact of settler-colonial risk factors [42].

Since its first release in 2018, analyses of the data collected have shown a positive relationship between connection to Country, culture, and health and well-being outcomes in relation to Aboriginal Ranger work in Central Australia [44]. Being employed as a Ranger, who uses cultural and environmental knowledge to engage in land management activities on Country, was significantly associated with very high life satisfaction and high family well-being [44]. Preliminary analyses of the full Mayi Kuwayu cohort show that key cultural indicators such as spending time on Country, speaking traditional languages, passing on family knowledge and traditions, and feeling in control of one’s life are protective against high psychological distress, diagnoses of anxiety, and low life satisfaction [45].

Mayi Kuwayu has also identified a range of settler-colonial exposures experienced by Aboriginal and Torres Strait Islander people and developed measures to capture participants’ exposure to these factors from the individual to the systemic level [39, 42]. These exposures have been classified as either Indigenous Historical Trauma (IHT) or Indigenous Contemporary Trauma (ICT). Indigenous Historical Trauma items included in the study are the following:

1. 1. Tribe/mob forcibly relocated to missions or reserves

2. 2. Being unsure of which tribe/mob you belong to

3. 3. Having a parent who was part of the Stolen Generations

4. 4. Having an aunt or uncle who was part of the Stolen Generations

5. 5. Having a grandparent or great-grandparent who was part of the Stolen Generations

ICT items include the following:

1. 1. Feeling disconnected from culture

2. 2. Being dislocated from Country

3. 3. Worrying about being stolen when they were growing up

4. 4. Growing up in foster care

5. 5. Growing up in a children’s home

6. 6. Having children removed in the past 12 months

7. 7. Experiencing interpersonal discrimination

8. 8. Being a part of the Stolen Generations

9. 9. Having a cousin who is part of the Stolen Generations

10. 10. Having child/ren who are part of the Stolen Generations

11. 11. Having grandchild/ren who are part of the Stolen Generations

Over 60 per cent of Mayi Kuwayu participants report at least one exposure to IHT and 85 per cent report at least one exposure to ICT [46]. Exposure to IHT is associated with significant increases in poorer psychological distress and poorer life satisfaction. Stronger links are observed between any experience of ICT and poorer psychological outcomes, poorer general health, and lower life satisfaction. More than half of the Mayi Kuwayu participants have experienced discrimination in some form, and this is significantly associated with a broad range of poor well-being outcomes, ranging from disconnection from culture to high blood pressure and alcohol dependence [47].

These findings strongly support Hoke and McDade’s [1] argument that studies on the DOHaD must consider the broader contexts in which we live as well as the intergenerational and transgenerational events that impact our health and well-being. Centring Indigenous lifeworlds in this research has the extraordinary potential to reveal previously ‘hidden’ factors that either increase the risk for disease or support good health and protect against biosocial harm. With the understanding that such factors and their contribution to health and well-being are heterogeneous across populations, more investment could be made by research institutions and funders in training DOHaD researchers to develop measures for specific salutogenic or risk factors in diverse populations. Intergenerational and transgenerational studies also require secure, long-term funding to provide more robust evidence as it accrues across time.

28.4 Conclusion

The world is currently undergoing a period of disruptive change driven by advances in data science and the convergence of technologies with the potential to enhance and harm Indigenous populations through analytics practice, including cohort studies. This can be addressed by ensuring Indigenous peoples are at the forefront of designing metrics and analytical studies. Data analytics and the translation of resulting insights into practice are key transformations affecting the future of society and its myriad of cultures. These innovations are a double-edged sword for Indigenous peoples, creating potential opportunities to improve well-being through the delivery of healthcare insights through a digital infrastructure that centres Indigenous values and protocols, but also raising concerns about data misuse and collective harm. Further, longitudinal cohort studies are a cornerstone of epidemiology and are central to knowledge production in DOHaD, but few have been developed by and for Indigenous peoples.

This chapter has argued that it is essential that contextual factors, including intergenerational and transgenerational factors, be accounted for by DOHaD research. Centring Indigenous lifeworlds has the extraordinary potential to identify previously ‘hidden’ factors and lead to the development of lifecourse interventions that could simultaneously reduce risks and increase protective factors. Through the Mayi Kuwayu study, for the first time in Australia and internationally, we have robust, national, longitudinal data on exposure to settler-colonial factors at the individual to systemic level as well as data on cultural practice and expression that may buffer the effects of settler-colonialism. Incorporating IDS and IDG frameworks and applying an Indigenous lens in the DOHaD research space has the potential to produce better science, better data, and better outcomes for our communities. Mayi Kuwayu is just one example of how Indigenous lifeworlds can be centred to produce a good data story about the origins of health and disease.

Chapter 29 DOHaD in the Anthropocene Taking Responsibility for Anthropogenic Biologies

Jörg Niewöhner

29.1 Entering the Anthropocene

As this volume has shown, the ‘developmental origins of health and disease’ framework (DOHaD) inquires into health and disease in adult human life as a function of environmental factors acting upon the human organism prior to conception, ante-/perinatally, in early life, and increasingly also throughout the lifecourse, respectively [1]. In its current form, which has been developed over the last three decades, it has not only helped to address the temporal and environmental dimensions of human disease aetiologies. This predominantly biomedical – in the broad sense of the term – framework has also proved useful to the social sciences and humanities to think through questions of human–environment relations, embodiment, and the role of the material environment in understanding ‘development’. The preceding sections in this volume attest to this generative role. They demonstrate how the framework acts as a boundary object, that is how it mediates between distinct disciplinary cultures. It also, however, carefully sensitises scholars to significant theoretical commitments, implicit assumptions, and practical consequences of current research on DOHaD conducted in these different disciplinary traditions. Many of these commitments are neither universally nor uncritically shared across academic disciplines. Attending to these differences is an important process in the development of DOHaD research.

In this final contribution to the volume, I want to look ahead and provide some tentative ideas about how the DOHaD framework might be translated into the Anthropocene. I understand the Anthropocene as the geological epoch following the Holocene and characterised by the acknowledgement that human action has developed into a formidable force shaping the planet in its entirety. More specifically, human action has been structured by the world’s dominant political economies into patterns of living and working that put immense pressure on the planet. So-called ‘planetary boundaries’ have been calculated that help to make visible how the planet responds [2]. These boundaries, such as climate, land use, biodiversity, ocean acidification, or the abundance of novel entities, have reached a point where the earth system might shift into radically different states that are likely to be far less amenable to human and social life than provided by the Holocene. Systems-speak aside, what this means is that deeply Western modern assumptions about progress, growth, and the stability of social expectations cease to exist as the unquestioned bedrock underpinning development and social welfare.

Today, tomorrow is less likely to be like yesterday. Instead, we are entering a phase of new extremes, new volatilities, and new non-linear dynamics and tipping points [3]. The Anthropocene challenges social and natural scientists to always also think in planetary terms. The key distinction between local and global, which has shaped academia and politics for decades, needs rethinking. The ‘terrestrial’ has instead been suggested as a way of conceptualising all social-ecological action in planetary terms [4]. Whichever way one frames the issue, one key question remains: how can societies worldwide establish and sustain this planet such that humans and other species can continue to inhabit it? Or as anthropologists today phrase it: how can more-than-human liveability on this planet be achieved? [5] The notion of ‘liveability’ indicates that this is not only a question of biological survival for human societies and beyond. It is a question of what a decent life can be. Hence, it is fundamentally a political and ethical question that is today reposed under conditions of rapid planetary environmental change and resultant social-ecological struggle and suffering.

Addressing the challenges of more-than-human liveability is a vast field. I want to focus here on one aspect only, namely the need to understand biology as anthropogenic [6]. By that I mean that both human bodies and the environments they inhabit are deeply shaped by human actions and the political economies within which these actions are organised. The steep rise of non-communicable, often chronic, as well as infectious diseases and mental health concerns correlates in astonishing ways with Western industrialisation and the global rise of capitalist means of organising human co-existence [7]. Natural resource exploitation, expanding industrial production, and the creation and mass production of novel biological and chemical entities at an unprecedented rate characterise this period of unchecked progress and development [8]. The result is landscapes, bodies, and metabolisms that are shaped by the dominant patterns of economic exchange and their violent histories of colonial extraction. These are landscapes, bodies, and metabolisms that can be meaningfully understood only as human-made: anthropogenic biology.

How then can the DOHaD framework address anthropogenic biology? How can it contribute to more-than-human liveability on this planet and to emerging thinking and research on planetary health? I offer my tentative line of argument in three steps: first, I want to make more graspable the challenges of more-than-human liveability to the DOHaD framework. To do so, I briefly discuss developments in environmental epigenetics, microbiome research, and the emergence of planetary boundaries. All three developments demand cooperation between natural and social sciences. I believe that the DOHaD framework currently does not offer a unifying solution as to how to organise this cooperation. Hence in a second step, I outline three different modes of interdisciplinarity between natural and social science following the excellent work of Andrew Barry and Georgina Born [9]: service, integration, and agonism. In a final third step, I set out what I consider important research questions and perspectives in each of these three modes that address more-than-human liveability while retaining the key concerns of the DOHaD framework. I conclude that the DOHaD framework must take responsibility for the emerging politics of habitability.

29.2 Rethinking Origins and Development in DOHaD in the Anthropocene

In this section, I reflect on recent research on epigenetics, the human microbiome, and planetary boundaries to draw out the implications for the notions of ‘origins’ and ‘development’ in the DOHaD framework.

29.2.1 Environmental Epigenetics

Environmental epigenetics denotes the study of changes in gene expression that occur without changes in DNA sequence [10]. Such changes may occur in response to environmental challenges to an organism, including both material (e.g. nutrients and toxicants) and social factors (e.g. discrimination and adversity). They operate through a number of mechanisms. The three most important currently known are methylation, histone- and RNA- modifications. Some of these changes have shown to be mitotically and meiotically heritable, that is they may propagate across generations. Many of the specifics of this field and its implications for DOHaD have been discussed elsewhere in this volume. I therefore keep this point brief.

Epigenetic research challenges the developmental origins of Western biomedical thinking in Mendelian genetics, Weissmann’s germ plasm theory, and – more broadly – the autonomous subject of Western modernity. If indeed heritable changes in germline functionality occur without DNA sequence change, individual human organisms are far more open to their past and present environments than so far appreciated. The DOHaD framework has begun to embrace these findings as they suggest mechanisms for phenomena that have so far only been shown through correlations [1]. The ‘tracking’ of physiological parameters over time from early life into adulthood might in fact be encoded at least to some degree in epigenetic processes. The details and some implications of this have been debated intensely over the last two decades. I want to focus on three lessons that follow from an epigenetics-inflected DOHaD understanding and that have received somewhat less attention.

First, ‘environments’ are diverse, dynamic, and never innocent. From my own fieldwork in a molecular biology lab working on environmental epigenetics, I have learned that epigenetic response mechanisms are far more subtle than the dominant digital logic of knock-out genetic thinking gives credit for [11]. Oftentimes, the handling of rodents in experimental settings appears to produce stronger epigenetic changes than the actual substance whose epigenetic effects were under investigation. This demonstrates that it might well be very difficult to isolate individual ‘factors’ or ‘causes’ from a complex environment. Instead, material and social factors readily interact in manifold ways with an organismic epigenome (if that term even makes sense), which is in itself a highly dynamic system. Mental models of human–environment relations derived from carcinogenicity and acute toxicity assume a unidirectional and non-reversible dose–response relationship between the environment and human organism (see Rossmann and Samaras in this volume). Epigenetics on the other hand suggests a reversible (e.g. de novo de/methylation) and perhaps bidirectional relationship where dose–response is likely to occur in a non-linear fashion; if indeed it is the right model at all. Lastly, environments are never politically and ethically innocent. The proof of principle experiments around the Dutch Winter Hunger cohort [12], the studies of licking and grooming behaviour under conditions of reduced nesting material and displacement [13], or the forensic psychiatric reconstruction of life histories in child abuse and suicide completers [14], all demonstrate that ‘environmental factors’ occur in politically, ethically, and socially charged settings that need to be appreciated in their economic, racial, and gendered complexity (see also the contributions by Meloni et al., Valdez and Lappé, and Cohen in this volume).

Second, while the notion of ‘origins’ in DOHaD was developed against genetic determinisms, it nevertheless still carries unwanted implicit remnants of Mendelian and Weissmannian biological temporality. Origins suggests a starting point that is defined if not genetically then still by some kind of non-human nature. This might not be a blank slate, but it is a starting point that is often reified through methodological designs that rarely reflect the constructed and contingent nature of study subjects – be they ready-to-study mice or Romanian orphans. The dynamic environmental conditions that have evolved historically and cross-generationally – from nutritional environments to political regimes – are seldom explored in social-ecological detail. Epigenetics makes us attentive to these conditions and thus to the historical and social contingency of any starting point. ‘All the way down’ [15] in the body, we do not encounter some kind of pure biological matter. Rather the body is as much ecosocially entangled at the molecular level as it is at the organismic level. For ‘environments’, the scientific construction work and the ecosocial entanglements are even more obvious. Finding a plausible starting point for one’s research design is a question of cutting the network and making the cut accountable to the field [16], one’s own discipline, and scientific practice at large. Epigenetics thus helps to challenge the idea of ‘origins’ as a largely unreflected, somehow natural starting point of a developmental process.

Third, for all the attention to environmental factors, the notion of development centres the framework on the human organism. That is perfectly acceptable as it is a framework in human medicine. Epigenetics, however, shows the human organism to be remarkably open to its manifold environments. And vice versa: the human organism contributes to making its own livelihoods and niches. Most ecologists today subscribe to the idea that organisms do not find or adapt to existing niches but that organisms and environments interact in co-producing niches [17]. In social scientific terms, ‘niching’ as a material and social everyday practice might be the more apt analytic for these forms of world-making [18]. Perhaps, then, one ought to refer to ‘genealogies’ instead of ‘origins’ and ‘development’. This would help to address the multiple histories that run through any origin as well as the necessary contingency of multiple struggles of power/knowledge that mark any genealogy. Genealogies of Health and Disease: GoHaD?

Appreciating the multi-directionality of human–environment relations casts doubt on the developmental thinking in environmental factors, mechanisms, and linear causality for all but the most pervasive and drastic isolated health effects. Most human–environment interaction research, however, remains rooted in a thinking premised on distinct entities. It is either enviro-centric suggesting that the environment as an independent variable causes certain responses in the body or it is organism-centric and thus focused on how human action shapes the environment [19] or how humans may act as niche modifiers [20]. These are all entity-based ways of thinking about human–environment interactions. They start from entities with certain characteristics (organisms and niches) that enter into interaction. One might also, however, start from the action and investigate how action produces entities. This results in a process-based approach [21]. Drawing on the French philosophers Gilles Deleuze and Félix Guattari, British anthropologist Tim Ingold proposes to think of humans in environments not in terms of interacting entities but as ‘lines of flight’[22]: ‘The line of flight, write Deleuze and Guattari, “is not defined by the points it connects, or by the points that compose it; on the contrary, it passes between points, it comes up the middle”’. What Ingold is essentially challenging his readers to do is think that humans and environments ‘should not be understood as interacting entities, … but as trajectories of movement, responding to each other in counterpoint, alternately as melody and refrain’. The result is a process- or practice-based biology in which organisms and environments are constantly in becoming and in which development occurs rhizomatically rather than along a linear path. The notion of development in DOHaD does not usually take this into account. It rests on the understanding of an entity that is exposed to an environment as a set of factors. I am not suggesting that lines of flight readily translate into biomedical research designs. Yet they present an important conceptual challenge that sensitises researchers to the fact that evolutionary, structural, and systemic thinking never quite captures the situational specificities of human practice and its effects. These require process-based approaches.

29.2.2 Microbiome Research

The human microbiome denotes the aggregate of all microorganisms living on or in human tissue or fluids. Research efforts to better understand the components and dynamics of the human microbiome have rapidly increased over the last decade. The human microbiome comprises around 10–100 trillion symbiotic microbial cells per human individual [23]. They match if not outnumber human somatic cells, and their genetic material by far exceeds that of the ‘human proper’. Cells belong to around 500–1000 different species at any given moment within a human [24]. The genetic diversity and hence flexibility of this crowd by far exceed that of human genetic material. Over the last ten years, the field of microbiome research has begun to transition from the description of components to mechanisms and to the tentative development of clinical interventions. It has also invited a rich scene of lay ‘bio hackers’ to self-experiment alongside the emerging science with everything from probiotic foodstuffs to faecal transfer. A scientific understanding of how the microbiome impacts human somatic and mental health and disease onset and progression directly, as well as through complex interactions with the immune, endocrine, and nervous system, is only just emerging. Yet already today it is becoming clear that microbiome research will be another insult to human narcissism and anthropocentrism. After Copernicus, Darwin, and Freud, microbiome research demonstrates that ‘man’ is not even somatically speaking the master in his own house. While the skin as ‘philosophy’s last line of defense’ [25] remains intact if porous, inside that skin shell emerges a multiplicity of inhabitants and agencies in complex and finely balanced interaction and oftentimes symbiosis.

In the preceding section, I used epigenetics to question whether we should move from origins to genealogies and from development to lines of flight. Microbiome research extends this questioning. Anthropologist Myra Hird discusses how human subjectivity and social form need to be understood as also shaped by bacteria [26]. She demonstrates how deeply biological and economic notions of the self are enshrined in Western modern thought. Westerners think of themselves as individual cognitive agents that act autonomously, often in competition with each other to increase fitness. Society is often understood as synchronised individuals. The vast majority of biological and medical research designs presume the existence of human individuals who act autonomously and are structurally closed to their environments. The subsequent distinction between self and other (along the skin) is foundational to Western self-understanding. Taking bacteria and their actions seriously challenges this understanding. Hird focuses on the understandings of symbiogenesis and a very corporeal interdependence of human and microbial life. The human ‘I’ becomes a multitude or collective. Similar to Ingold, she thus arrives at a world in becoming that is made up of encounters: encounters, for example, between humans and microorganisms that then develop ways of co-existing. Continuous encountering is what makes us what we are. Hence ‘we’ are not only epigenetically open to ‘outside’ environments, but ‘we’ are also open to ‘inside’ environments in the form of microorganismic collectives and their respective habitats.

It is this continuity of encounters that DOHaD also needs to address. Whether framed as contagion, as multi-species thinking, or as making kin, health and disease can rarely be sensibly understood as states of a single organism isolated in interaction from its environment. Dose–response simply does not capture the fact that exposure occurs in dynamic patterns of encounters. Even in the simplest scenario of an isolated single toxicant having an impact on a human organism, it is not only human cells and organs reacting. It is a symbiogenetically evolved social organism that responds. And while this social organism exhibits a distinct meta-stability that most people readily accept as human subjectivity, responses to substances are multiple and differentiated. Substances that occur at levels in the environment that are commonly considered well below toxicity thresholds not only interact to complicate exposure. They interact in differential ways with the human microbiome such that effects may occur that might well then surface as a human health issue. Hence environments are never really environments for only one organism as Uexküll suggested. The ‘environment multiple’ is a direct result of understanding the human body as a multiplicity of encounters.

29.2.3 Planetary Boundaries and Planetarity

So far, I have addressed molecular, cellular, and organismic dynamics. Let me now briefly turn to planetary dynamics and how they might challenge DOHaD. Earth system science is understanding with increasing certainty that the planet’s capacity to sustain life as we know it has boundaries that we are already transgressing through human action [2]. Environmental factors are thus ceasing to be local phenomena and instead need to be contextualised within planetary environmental change and its manifold repercussions for social-ecological systems across the globe. Philosopher Bruno Latour rightly challenges us to develop across all forms of scholarly activity a consciousness for the fragile and restricted conditions of habitability of the earth and life on it [4].

This planetary dimension challenges the DOHaD framework to understand ‘the environment’ and its health-relevant factors as part of earthly subsystems and its associated complexities and non-linear dynamics [20]. Major efforts have been underway for some years now to better understand and quantify both the global burden of disease [27] and the bio-geo-physical and increasingly social dynamics of earth’s subsystems as well as their stable state boundaries [2]. Suggestions are being made to integrate earth system science and global health through international consortia and (big) data approaches [28]. Such approaches are commonly shaped by systems thinking and various forms of computational modelling that foster data-driven integration. The rise of earth system modelling from the late 1980s onwards has undoubtedly been an amazing process of knowledge production culminating in the 6th Assessment Report of the IPCC on the ‘physical science basis’ of global climate change. Never has a comparable global evidence machine been built of such scale and with such rigour. This evidence machine runs on a positivist epistemology that addresses ‘the planet’ through data aggregation and integration as part of system dynamics modelling. Its dominant if not its only mode of speaking about large phenomena (aka ‘the planet’) is through scaling up by means of data aggregation and integration. Everything else is anecdotal or an opinion, that is not considered evidence. This approach aligns well with the data-driven calculations of global burdens of disease and exposomes.

Yet feminist literary theorist Gayatri Spivak [29] and others use the notion of ‘planetarity’ [30] to point to the data-driven construction of the planet asking where this leaves significant differences in ways of thinking and being on this planet – differences to which the social sciences and humanities attend. An altogether different perspective on this new planetary dimension is thus possible and important. The social sciences and humanities have long developed conceptual and empirical alternatives to address large-scale phenomena such as globalisation. Thinking in scapes and flows [31], global assemblages [32], global entanglement [33], and post- and decolonial critique [34] approaches ‘the global’ very differently. These approaches start from significant social differences and ask how they spread, reach, infect, travel, transform, and resonate. These approaches either focus on the forces that make phenomena graspable as ‘global’ or they move around inside the phenomena understood as global, showing their heterogeneity and multiplicity. Globality is about differences and what these differences (can) mean for respective others. Planetarity, in contrast, is often about trying to produce the one true representation of a whole. Of course, globality has been concerned with primarily social dynamics, while planetarity is rooted in biogeophysical phenomena. In the Anthropocene, however, this neat division of labour is being challenged as social inquiry becomes interested in material dynamics and ontological questions, while modellers of physical systems are incorporating not only economic exchanges into their models but increasingly also social dynamics.

Latour’s ‘Terrestrial’ [4] might be a useful point of contact between these very different ways of addressing phenomena that span the world in various ways. It is these reconfigurations of socio-material relations and the ensuing debate of how to address them that form the context within which the DOHaD framework can make an important contribution. How do we conceptualise ‘environmental factors’ when they cease to be local phenomena; when we try to understand them within a terrestrial context? One response would be to scale ‘environmental factors’ up, for example, to produce global estimates of ‘novel entities’, that is ‘new substances, new forms of existing substances and modified forms of life’ [8], and assess their impact on earth’s subsystems. Microplastics, lead, or persistent organic pollutants serve as examples. This might be related to the global burdens of disease in a data-driven integrative approach. In a very different approach, one might contextualise ‘environmental factors’ in highly political patterns of exposure related to colonial and racialised histories of exploitation and production, imperial debris, and the ruins of capitalism [35].

The Anthropocene and its demand to think in planetary terms, then, challenge biomedically and epidemiologically established notions of environment, environmental factors, exposure, dose–response, temporality, and scale. These notions are also embedded within the DOHaD framework. However, the framework does not inherently offer a singular and straight path to address these challenges. It is not a foregone conclusion that these challenges will be solved with data collected and analysed in the empirico-analytical frameworks of twentieth-century biomedicine. Rather, the DOHaD framework might afford a reflexive moment, a moment of producing ‘theory out of science’ [36], which might enable a diverse set of approaches to address anthropogenic challenges and to address more-than-human liveability on a fragile planet. The DOHaD framework might offer a space within which data-driven approaches might complement other approaches that open up environments to a politics of exposure and habitability and that understand the human body as historically shaped, socialised, and habituated in patterns of practice [37]. Realising this potential and situating DOHaD in this sense, however, requires a diversity of approaches.

29.3 Modes of Interdisciplinarity: Ecosocial Co-laboration

29.3.1 Three Modes of Interdisciplinarity

Moving DOHaD into the Anthropocene cannot be achieved with a single logic or methodological approach. It would be a futile task to try to develop an overarching framework that can fully integrate the existing conceptual and methodological diversity, the desire to reduce and explain with the need to contextualise and interpret, and the strengths of rigorous data-based analysis with the strengths of critical inquiry. The present volume features this diversity, and it is clear that this does not cohere in an overarching heuristic or integrative framework – nor should it. Instead, it might be useful to distinguish between different forms of interdisciplinary engagement and forms of collaboration to give orientation and take away some of the pressure towards integration.

Andrew Barry and Georgina Born [9], drawing on their investigation of using ethnography within various other epistemic cultures, usefully outline three modes of interdisciplinarity: subordination–service, integration–synthesis, and agonistic–antagonistic. In the subordination–service mode, the research question and design come from one lead discipline, while the other discipline delivers additional data to extend or deepen the analysis. This is a fundamentally asymmetrical approach shaped by one discipline. In the integration–synthesis mode, two different disciplines readily find ways of addressing problems that are of interest to both. Biochemistry is the typical example within the natural sciences. It is a little less obvious across the natural/social science divide. This approach is symmetrical with both disciplines staying within their comfort zone but addressing a new topic in a shared way [3]. The agonistic–antagonistic mode is perhaps the most demanding. It arises when two disciplines differ in their understanding of the research object. Often, this disagreement is ontological in nature. For example, human differences might be considered a material and bodily phenomenon by biologists while social scientists would insist on it being primarily a social phenomenon constructed through the social interaction of subjects positioned in social space. In an agonistic–antagonistic mode, these differences between disciplinary perspectives are not levelled. Instead, they need to be worked with to turn them into something generative from which both disciplines can learn without necessarily agreeing to the respective other perspective. This approach is symmetrical but not as comfortable as in the integration mode. It is about letting oneself be irritated by other ways of thinking and designing research and by sustaining significant differences to learn and develop one’s own perspective.

In opening up DOHaD further to social science thinking, this form of agonistic interdisciplinarity will play a key role. It is important to note that co-laboration [38] is possible: co-laboration is temporary joint knowledge production between two disciplines without necessarily having a shared goal. Two scholars from different disciplines might work together without necessarily getting a shared result. Rather they might take different results and insights away from the co-laboration and integrate those into their respective disciplines. To make this rather abstract typology of possible interdisciplinary research more tangible, I want to briefly sketch examples of research questions and designs for each of these three modes of interdisciplinarity that might help the current DOHaD framework embrace the challenges of the Anthropocene.

29.3.2 Subordination–Service

Environmental epigenetics offers an obvious entry point for subordination–service interdisciplinarity. Currently, research within the DOHaD framework tends to operationalise environmental factors as independent variables, for example social disadvantage, zip code, and nutritional status (cf. Liz Roberts in this volume). For many social situations and structural inequalities [39], such operationalisations are not only too crude. They also miss the entire dimension of subjective and collective experience that social science would consider fundamental to the emergence of ‘the social’ and thus to pathways from ‘objective’ disadvantage to actually ‘living inferiority’ [39] and associated individual bodily health and disease. Hence social science could contribute its understanding of social dynamics to research on social drivers within the DOHaD framework. At a time where many countries – and in particular the major metropoles – are undergoing fundamental transformations trying to meet climate targets, this kind of social science service work could also contribute to making sure that ambitious climate targets are not reached at the expense of increasing inequality and thus worsening individual and public health outcomes.

Subordination–service approaches might also work the other way around. The long-term social inquiry into living with chronic respiratory diseases, for example, might benefit from global to local climate projections, air quality modelling, and associated health data. The Anthropocene foregrounds the dynamics of nature–culture relations, making such social inquiries into ‘natural’ phenomena such as disease aetiologies even more relevant [40]. Planetary health is currently conceptualised largely in biomedical and public health terms with the social sciences adding knowledge about social dynamics. This could be – perhaps ought to be – turned around placing well-being and environmental justice at the centre and putting the medical disciplines in the service role.

29.3.3 Integration–Synthesis

Recent work at the interface of sociology and biogeochemistry presents an outstanding example of biosocial interdisciplinarity in the integration–synthesis mode [6]. Hannah Landecker, a sociologist and historian of science, and Cajetan Neubauer, a biochemist, in rather serendipitous fashion, began to work on the planetary availability and bodily effects of methionine together. Methionine is an essential amino acid, that is it cannot be synthesised by the human body and needs to be taken up through food. Working together on the methionine metabolism both globally and within the body, the two perspectives together were able to ‘establish the scale and historical trajectory of the methionine industry and provide a preliminary model for tracing this amino acid through the food supply into the human body’ [6]. The study shows how planetary-scale anthropogenic activity changes ‘environments’ and consequently also human metabolisms with so far largely unknown consequences. Human biology and ecology, that is both environment and body, are understood as anthropogenic. The DOHaD framework thus requires biosocial, or rather ecosocial, collaboration as it is the social science perspective that can reveal how environmental factors come to be what they are through anthropogenic activity, specifically through analyses of dominant political economies and ecologies.

Hannah Landecker reflects on the back of the methionine study: ‘I have always felt that my contribution has been to enable the asking of experimental questions, the parameterisation of models, and the forming of hypotheses that would not otherwise have been possible … What Harry Collins termed “interactional expertise” has also been important in helping teams of different kinds of knowledge practitioners recognise ways in which they don’t understand one another, or facilitating the synthesis of different modes of proof or reasoning….’ [41] Landecker reports her interdisciplinary research as an example of integration–synthesis cooperation, that is both disciplines, biochemistry and history, contribute from within their comfort zone to arrive at a new question and analysis. This is cleverly done, and it is this approach that enables Landecker to conduct a historical and social study of chemicals of metabolic significance – as a social scientist, albeit one with deep knowledge of natural science and the dynamics of the particular field in question.

29.3.4 Agonism–Antagonism

Oftentimes, however, social science perspectives do not readily complement or integrate with existing natural science or medical knowledge. Take approaches to social dynamics as an example. Biomedical or public health operationalisations of social dynamics often do not resonate with state-of-the-art social scientific knowledge and critique. The reason for such dissonance often lies in profound differences in the understanding of the research object. Much medical expertise works with a concept of ‘the social’ that is based on individuals interacting within a value system or culture. The notion of the individual tends to be under-socialised, that is based on ideas of individual decision-making that might be found in behavioural economics. Whereas the notion of the ‘value system’ tends to be over-socialised, that is assuming a firm grip of an abstract but homogeneous set of values on the framing of individual behaviour. Social dynamics from most social science perspectives sit in between these two perspectives and often work with notions of agency, subjectivity, and practice that combine structural (‘value system’) and individual (‘decision-making’) elements in highly dynamic and reflexive ways.

The agonistic–antagonistic approach starts from such dissonances and tries to make them generative. Agonism here refers to the work of political scientist Chantal Mouffe, who argued for agonism as a democratic form that does not solely rely on consensus but rather is able to work with differences as a potentially positive democratic force. Agonism within interdisciplinary research then means working with and through differences rather than searching for an integrative framework. This takes some work between the right partners. Working with differences requires explicating methodological, epistemological, and ontological assumptions. Differences need to come out to enable research that searches for common ground while critically reflecting on one’s own disciplinary perspective.

In this regard, the DOHaD framework needs to address the question of how much it can allow environments and bodies to be ‘situated’ in a social science sense [42]. As it stands, DOHaD often rests on a universal body that responds to environmental factors. While this rests on a perfectly plausible set of assumptions, it does appear to underestimate what anthropogenic biology means. If all-pervasive anthropogenic activity has begun to change environments and bodies in significant ways, the ‘universal’ body is perhaps not the most prudent assumption anymore. Of course, at some general level, most bodies share some basic structures and processes. Yet these structures and processes are shaped by widely different trajectories through individual life histories and start to differentiate in significant ways. Most study designs assume a universal body and investigate the response to exposure. Yet ‘exposure’ to chemicals, social disadvantage, infrastructural violence, and to colonial legacies is systematically patterned and has been so for many centuries. The ‘inward laboratory’ of the body adjusts and contributes to these patterns and to living in, with, and through these patterns in significant ways. Rather than starting from the universal body at a national or even global scale, is it not high time that we should also try to scale approaches a little differently? Sure, trying to assess the global burden of disease, the global presence of chemicals and novel entities, and planetary boundaries remains important. Systems thinking and data-based integration remain important, and attempts at measuring the exposome and conducting exposome-wide association studies will certainly be productive in many ways even if they are bound to never reach their objective [43]. Yet other exposure patterns also matter. We might start with landscapes and how they have been shaped by political histories and economies, social dynamics and forms of dwelling, as well as biogeophysical contexts of climate, topography, and ecology. Urban and rural landscapes [5] can be investigated in great ecosocial detail to better understand how they afford particular exposure patterns and how these contribute to shaping deeply situated bodies [44]. Inside such landscapes, situated bodies are thoroughly historicised, socialised, and politicised and constantly in becoming. They are not simply local as they relate to many transnational exchanges and flows of people, goods, and information as well as planetary environmental change. Such ecosocial analyses of the situated genealogies of health and disease require the agonistic and antagonistic struggle between critical social science perspectives and rather more solution-oriented medical perspectives.

29.4 The Future of DOHaD: Taking Responsibility for Anthropogenic Biologies

Translating the DOHaD framework into the Anthropocene requires an opening up of its underlying biomedical and epidemiological thought style. It needs to be opened up, because the human body and the complex social-ecological environments within which it dwells can only be understood as anthropogenic. They have ceased to be ‘natural’ in any meaningful sense. In an epoch where human action and its political economy are thoroughly transforming life at a planetary scale, the material agencies of more-than-human liveability cease to be universal and innocent – if they ever were. Situating bodies and environments is one response. Situating means understanding how ‘environmental factors’ are embedded within landscapes that are shaped thoroughly by land use practices and metabolisms reflecting dominant natural, social, and moral orders. And it means appreciating the habituated nature of the human body-in-practice that fends for its livelihood in such landscapes.

The DOHaD framework offers sufficient openness to pursue health and disease in emerging and dynamic patterns of everyday practice. The key to such an approach is the constant careful calibration of a balance between on the one hand appreciating the singularity of life as such [45] and its multi-species encounters and on the other recognising structural continuities in life itself [46] shaped by hegemonic patterns of bio- and geopolitics regulating bodies and landscapes. Methodologically, this has to be a programme that starts from an in-depth understanding of the regularities and patterns that shape health and disease over time in situated cases. Long-term social-ecological field sites that can understand exposure as practice rather than correlation seem promising. Ethnographic, micro-sociological, and micro-historical approaches need to be brought into conversation with, on the one hand, biomedical and epidemiological methods and, on the other, with methods that can assess drivers of landscape change, for example earth observation, land use science, and climate impact modelling. Such multi-method approaches need not be fully integrated into a single framework. They can explore the tension between the singularities and regularities of more-than-human liveability by exploring both statistical and analytical generalisation; by situating numerical models of social-ecological dynamics; and by deconstructing knowledge claims but also by reconstructing alternatives.

Starting from situated cases and possibly community-based research into more-than-human liveability also offers the chance to embrace the necessarily political nature of this work. Long-term multi-method field sites offer the opportunity to align knowledge production with the co-production of interventions rooted in a thoroughly situated understanding of the developmental origins of multi-species health and disease. In such an approach, the DOHaD framework begins to take responsibility for how bodies and environments are known and problematised. It begins to take responsibility for the worlds people live in and for possible futures.